Experiment · data coverage · figures · study design — Zhong et al. 2025

How neuronal selectivity distributions change with experience

Source-linked map of the published results, complete release inventory, data provenance, and two explicitly identified proposed analyses of selectivity distributions and within-session dynamics.

Nature 644, 741 · 2025 19 imaging mice · 89 recordings 20,547–89,577 Suite2p traces/recording 421.2 GiB · 297 files CC BY 4.0

Every paper-derived claim links to the corresponding Nature article or figure page. Release-audit numbers link to the article’s deposited dataset, and the open full text, analysis code, and methods review remain one click away.

Nature article Nature PDF in Drive Open-access full text Figshare v2 data Released analysis code Neural-data methods review Methods PDF in Drive
01
Nature paper

Unsupervised pretraining in biological neural networks

The paper reports 89 recordings in 19 TetO-GCaMP6s × CaMK2a-tTA mice (Methods: animals ↗) and 20,547–89,577 Suite2p-derived traces per recording across V1 and higher visual areas (Results: supervised and unsupervised plasticity ↗; Figure 1e ↗). The rewarded task and unrewarded-exposure cohorts are defined in the same Results section and timeline (Results section ↗; Figure 1a–b ↗).

After Train 1, the paper reports increased familiar-stimulus selectivity in task and unrewarded natural-texture cohorts, with no comparable increase after grating exposure (Results: supervised and unsupervised plasticity ↗; Figure 1i–j ↗). Separately, it reports a late-versus-early cue-position response concentrated primarily in anterior visual areas of task mice after training (Results: reward prediction in anterior HVAs ↗; Figure 4e–g ↗).

Primary sources and exact figure pages
Publication, data, code, methods, and figure links used in this reference
SourceDirect linksScope in this reference
PublicationMain text ↗ · Methods ↗ · Statistics and reproducibility ↗ · Data availability ↗Each claim below links to a narrower figure, Results section, or Methods subsection.
Main figuresFig. 1paper ↗ · Fig. 2paper ↗ · Fig. 3paper ↗ · Fig. 4paper ↗ · Fig. 5paper ↗Panel-level interpretation and reproduction map
Extended DataED Fig. 1paper ↗ · ED Fig. 2paper ↗ · ED Fig. 3paper ↗ · ED Fig. 4paper ↗ · ED Fig. 5paper ↗ · ED Fig. 6paper ↗ · ED Fig. 7paper ↗ · ED Fig. 8paper ↗ · ED Fig. 9paper ↗Complete supporting figures, placed beside the corresponding main result.
Data and codeFigshare v2 · paper-tagged codeReleased files and executable figure recipes
Analysis methodsStringer & Pachitariu 2024 · open PDFCross-validation, controls, and interpretation safeguards

The paper reports before-versus-after selectivity and cortical-distribution endpoints at the Train 1 landmarks (Figure 1b,i,jpaper ↗). In the deposited Imaging_Exp_info.npy experiment index ↗, matched Train 1 acquisitions are 3–24 days apart overall (supervised: 9–24 days; unrewarded: 3–12 days). Figure 1i–j summarizes before/after selective-neuron fractions rather than chronological within-session d′ distributions; that separate proposed analysis is defined in Research question 1.

Main Figures 1–5

Complete published figures are shown once, in numerical order. Panel summaries below paraphrase the corresponding Nature captions; each scientific statement links to the exact Results, Methods or figure location.

Figure source. Figures 1–5 and Extended Data Figures 1–9 are reproduced unchanged from Zhong et al., Nature 644, 741–748 (2025) under CC BY 4.0. Every caption links to its exact figure page.

01

Nature Figure 1

Plasticity in the visual cortex after supervised and unsupervised training.

Nature Figure 1 showing the task, training cohorts, behavior, imaging, selectivity index, neural responses, cortical maps, and regional fractions
Figure 1 — Plasticity in the visual cortex after supervised and unsupervised training. Panels a–j report the task and cohort timeline, anticipatory licking, mesoscope imaging, the signed selectivity index, example selective responses, cortical density maps and regional selective-neuron fractions. Exact figure on Nature ↗
Extended Data Figure 1 showing stimulus crops, retinotopy, threshold sweeps, signed selective-neuron fractions, and V1 subdivisions
Extended Data Figure 1 — Retinotopy and neural changes after learning for different populations. Panels a–g show example stimulus crops, retinotopic mapping and atlas alignment, the grating-cohort selective-neuron distribution, threshold sweeps, separate leaf1- and circle1-selective fractions, and medial-versus-lateral V1 summaries. Exact figure on Nature ↗
Extended Data Figure 2 showing running speed and percentage of time running before and after learning
Extended Data Figure 2 — Running behaviors. Panels a–d show position-resolved running speed in example unrewarded and task mice, before-versus-after speed summaries, and percentage of time running. Exact figure on Nature ↗
Figure 1a–b ↗
The virtual-reality corridors contain a random-position sound cue; water is available after the cue in the rewarded corridor. The timeline distinguishes task, unrewarded natural-texture and unrewarded grating cohorts.
Figure 1c–d ↗
An example lick raster and anticipatory-lick summary report task behavior after learning.
Figure 1e–f ↗
Mesoscope coverage and cellular resolution are shown beside the signed leaf1-versus-circle1 d′ definition.
Figure 1g–h ↗
Single-trial responses are shown for example circle1-selective and leaf1-selective neurons and their corresponding populations.
Figure 1i–j ↗
Aligned cortical density maps and regional fractions compare selective neurons before and after learning or exposure.
02

Nature Figure 2

Comparing visual and spatial coding on test stimuli.

Nature Figure 2 showing test stimuli, licking, neural sequence comparisons, coding-direction projections, and similarity indices
Figure 2 — Comparing visual and spatial coding on test stimuli. Panels a–j report test 1 stimuli and licking, preferred-position sequence comparisons, leaf1- and circle1-selective populations, coding-direction projections and similarity indices. Exact figure on Nature ↗
Extended Data Figure 3 showing circle1-selective sequences and coding-direction projections across cohorts
Extended Data Figure 3 — Sequences in circle1-preferring neurons and average projections on the coding direction. Panels a–e show circle1-selective sequence analyses and coding-direction projections for naive, unrewarded and task cohorts. Exact figure on Nature ↗
Figure 2a–c ↗
Test 1 presents familiar leaf1 and circle1 corridors with new leaf2 and circle2 exemplars; the figure shows an example lick raster and the five-mouse anticipatory-lick summary.
Figure 2d–f ↗
Responses are sorted by preferred leaf1 position on held-out trials; preferred-position correlations are summarized within task mice and across regions and cohorts.
Figure 2g–j ↗
Leaf1- and circle1-selective populations define a coding direction. Test-trial projections and a similarity index summarize the four stimuli across regions and cohorts.
03

Nature Figure 3

Responses to novel and adapted stimuli and neural orthogonalization.

Nature Figure 3 showing novel and adapted stimulus responses, licking, leaf1-versus-leaf2 selectivity, projections, and orthogonalization
Figure 3 — Responses to novel and adapted stimuli and neural orthogonalization. Panels a–h report leaf2 novelty and adaptation, leaf1-versus-leaf2 selectivity, licking, coding-direction projections and similarity indices. Exact figure on Nature ↗
Extended Data Figure 5 showing circle2-selective and leaf1-selective neuron results
Extended Data Figure 5 — Representations of familiar versus novel stimuli. Panels a–b show circle2-selective-neuron results corresponding to Figure 3a and leaf1-selective-neuron results corresponding to Figure 3b. Exact figure on Nature ↗
Extended Data Figure 6 showing leaf3 recognition, de-orthogonalization, and leaf1-swap analyses
Extended Data Figure 6 — Visual recognition memory becomes exemplar-specific after extended training. Panels a–i show leaf3 behavior and projections on the leaf1–leaf2 axis, the reported de-orthogonalization schematic, and leaf1-swap behavioral and neural analyses. Exact figure on Nature ↗
Extended Data Figure 7 showing test 3 swap stimuli, coding-direction projections, similarity indices, and licking
Extended Data Figure 7 — Coding direction projections and similarity indices during test3. Panels a–g show a second swap stimulus, V1 and medial coding-direction examples, similarity indices, and test 3 licking summaries. Exact figure on Nature ↗
Figure 3a–b ↗
Leaf2-versus-circle1 selective-neuron distributions and regional summaries compare leaf2 when new and after learning, with task, unrewarded and naive observations.
Figure 3c ↗
Licking to leaf2 is compared when it is new and after task training.
Figure 3d–e ↗
Leaf1-versus-leaf2 selective-neuron distributions and regional fractions compare task, unrewarded, naive and grating-control observations.
Figure 3f–h ↗
Leaf2 responses are projected onto the leaf1–circle1 coding direction in V1 and the medial region; the similarity index is summarized across regions and cohorts.
04

Nature Figure 4

A reward-prediction signal in supervised training only.

Nature Figure 4 showing Rastermap responses, reward-prediction neuron locations, cue-position selectivity, and behavioral controls
Figure 4 — A reward-prediction signal in supervised training only. Panels a–n report Rastermap-guided discovery, cortical location, a late-versus-early cue-position index, regional distributions and response controls across test sessions. Exact figure on Nature ↗
Extended Data Figure 4 showing pre-cue licking, coding-direction projections, and regional lick-versus-no-lick comparisons
Extended Data Figure 4 — Relation between neural activity and licking behaviors. Panels a–d show pre-cue licking on early- and late-cue trials, coding-direction projections on those trials, and the lick-versus-no-lick comparison corresponding to Figure 4l across regions. Exact figure on Nature ↗
Extended Data Figure 8 showing reward and non-reward prediction neurons, running, and licking across areas
Extended Data Figure 8 — Reward and non-reward prediction neurons across areas. Panels a–f show reward-prediction fractions by region, circle1-defined non-reward-prediction controls, and cue-aligned neural activity, running speed and licking rate across four corridors. Exact figure on Nature ↗
Figure 4a–c ↗
A Rastermap-ordered population, an enlarged leaf1-active cluster and the selected cells’ cortical positions are shown.
Figure 4d–g ↗
Leaf1 trials are split by sound-cue position to define late-versus-early cue-position d′; selected-neuron distributions and anterior-region fractions are compared before and after learning.
Figure 4h–l ↗
Test 1 responses are shown across corridors, aligned to cue or first lick, and split by lick versus no-lick trials.
Figure 4m–n ↗
Population responses are shown in test 2 and test 3.
05

Nature Figure 5

Unsupervised pretraining accelerates subsequent task learning.

Nature Figure 5 showing pretraining cohorts, task structure, licking, learning curves, first-lick locations, and trial counts
Figure 5 — Unsupervised pretraining accelerates subsequent task learning. Panels a–h report the three pretraining cohorts, task design, example licking, group learning curves, first-lick locations and daily trial counts. Exact figure on Nature ↗
Extended Data Figure 9 showing within-day behavioral performance by pretraining cohort
Extended Data Figure 9 — Within-day learning in unsupervised pretraining experiment. Behavioral performance is split into the first and second halves of each session for 11 naturalistic-pretrained, 7 grating-pretrained and 5 no-pretraining mice. Exact figure on Nature ↗
Figure 5a–b ↗
Naturalistic-texture, grating and no-pretraining cohorts undergo five task-training days. The sound cue is absent; reward is available in the second half of the rewarded corridor, contingent on licking except during the initial passive-reward day.
Figure 5c–d ↗
Example lick distributions are shown for the first active-reward day and the last training day.
Figure 5e–f ↗
Mean lick responses and rewarded-minus-non-rewarded performance are reported across days for 11 naturalistic-pretrained, 7 grating-pretrained and 5 no-pretraining mice.
Figure 5g–h ↗
First-lick locations and daily trial counts are reported for the three cohorts.
M

Nature Methods

Acquisition, processing and observation units

Imaging acquisition ↗
Two-photon mesoscope recordings sampled multiple visual areas simultaneously.
Calcium processing ↗
The paper used Suite2p-derived deconvolved activity with a 0.75 s decay parameter.
Neural selectivity ↗
The selectivity calculation retained original, non-interpolated running frames in the 0–4 m textured segment and defined signed d′ from the two corridor-response distributions.
Retinotopy ↗
Recording-specific retinotopic maps were aligned to a reference mouse and used to assign cortical regions.
Statistics and reproducibility ↗
Panel captions state the mouse or session counts and whether paired or independent two-sided Student’s t-tests were used.
02
Mice, cohorts & stages

The animals and the training schedule

The paper’s Animals Methods report 89 recordings in 19 TetO-GCaMP6s × CaMK2a-tTA mice (13 male, 6 female; 2–11 months) (Methods: animals ↗). The filename key recording_id = {mouse}_{date}_{block} and cohort prefixes come from the deposited Imaging_Exp_info.npy ↗, not from the prose paper. Counts are panel-specific: Figure 1j ↗ reports n=4 task, 9 unrewarded natural-texture, and 3 grating mice; other panels use the counts shown explicitly below.

Imaging cohorts and reward conditions
Paper cohortRelease prefixReward protocolPanel-specific nExact evidence
Tasksup_*Sound cue marks reward-zone onset; water follows a post-cue lick in the rewarded corridor, with passive-delay variants in some mice.n=4 in Fig. 1j; n=5 in Fig. 2e,f,j.Reward protocol ↗; Figure 1j sample size ↗; Figure 2e,f,j sample sizes ↗; prefix and rewType rows ↗
Unrewardedunsup_*Same imaging corridors and sound cue; no rewards.n=9 in Fig. 1j; n=7 in Fig. 2f; n=6 in Fig. 3b,e,h.Unrewarded protocol ↗; Figure 1j sample size ↗; Figure 2f sample size ↗; Figure 3b,e,h sample sizes ↗; release prefix rows ↗
Grating control*_gratingUnrewarded 0°/45° grating exposure; neural responses are tested on naturalistic pairs before and after.n=3 (5 sessions) in Fig. 1j; n=3 (6 sessions) in Fig. 3e,h.Grating and naturalistic stimulus protocol ↗; Figure 1j sample size ↗; Figure 3e,h sample sizes ↗; release prefix rows ↗
Naivenaive_*No training or exposure; more than one naturalistic pair can be imaged per mouse.n=9 mice (11 sessions) in Fig. 2f and Fig. 3e,h.Naive stimulus protocol ↗; Figure 2f sample size ↗; Figure 3e,h sample sizes ↗; release prefix rows ↗
A distinct behaviour-only cohort

Fig. 5paper ↗ and Extended Data Fig. 9paper ↗ use 23 additional mice with no imaging, split 11 naturalistic-pretrained / 7 grating-pretrained / 5 no-pretrain, followed by a simplified 5-day reward task. Their behaviour files use a per-trial schema (LickPos, LickTrind, WallType) keyed {mouse}_{date}_dayN, distinct from the imaging per-frame ft_* arrays, and carry no neural data.

Stage schedule

The task and unrewarded imaging cohorts follow the staged Train/Test sequence shown in Figure 1bpaper ↗. The deposited Imaging_Exp_info.npy experiment index ↗ gives matched Train 1 acquisition gaps of 3–24 days overall (supervised: 9–24 days; unrewarded: 3–12 days). Naive and grating controls contribute comparison sessions rather than passing through every task stage; the release index contains 23 experiment labels spanning cohort, stage, and learning status.

Experimental stages and released file families
StageDurationStimuliDescription · files
Acclimation>8 daysgrey / spoutHead-fix and running habituation (>6 cm/s); restricted mice receive spout training.
Train 1~2 weekscircle1 (0) vs leaf1 (2)Coarse discrimination; task mice lick selectively to leaf1. Beh_{sup,unsup}_train1_{before,after}_learning.npy, {rec}_SVD_dec.npy, {rec}_trans.npz. These matched milestones define the four study strata in §11.
Test 1test session+circle2 (1), leaf2 (3)New exemplars; licking generalises to leaf2, not circle2. Basis of the visual-versus-spatial analysis (Fig. 2paper ↗).
Train 2~1 weekcircle1, leaf1, leaf2Fine-discrimination training preceded Test 2. After training, the leaf2 representation had a smaller projection onto the leaf1–circle1 coding direction, with the largest reported change in medial HVAs (Results: novelty and orthogonalization ↗; Figure 3f–h ↗).
Test 2test session+leaf3 (4)Novel exemplar; coding-direction generalisation.
Test 3test session+leaf1_swap (5,6)Spatially rearranged leaf1; tests stimulus identity against position.
Methods detail: corridor, cue, reward, movement, and cohort differences
  • The imaging corridor contained 4 m of pseudorandom naturalistic texture followed by 2 m of grey (Methods: visual stimuli ↗). The sound cue was sampled from 0.5 to 3.5 m in every task and unrewarded imaging trial, while water could follow it only in the rewarded corridor for task mice (Methods: behavioural training ↗).
  • Circle, leaf, rock, and brick are physical texture families in the paper (Methods: visual stimuli ↗); stim_id and the released Imaging_Exp_info.npy provide the session-specific role mapping (exact released index file ↗).
  • Train 1 lasted about two weeks, Test 1 introduced new exemplars, Train 2 lasted about one week, Test 2 introduced leaf3, and Test 3 rearranged the familiar texture spatially (Figure 1b timeline ↗).
  • The virtual corridor moved at 60 cm s−1 while running exceeded 6 cm s−1 (Methods: visual stimuli ↗); the paper’s selectivity calculation then retained running, non-interpolated frames within the 0–4 m textured segment (Methods: neural selectivity ↗). The broader speed/occupancy/cue/reward/lick QC set is a proposed safeguard here, not a statement that the paper fit all of those variables jointly.
  • The paper states that some imaging mice received passive reward after a delay (Methods: behavioural training ↗); the exact session-level rewType variants and the 4 rewarded versus 9 unrewarded Train 1-before rows come from the released index (exact released index file ↗; rendered release audit ↓). Selecting that comparison for RQ2 is a proposed design choice.
03
Stimuli

Stimulus roles are session-specific

The paper refers to stimulus roles in its naturalistic-corridor design (Nature Fig. 1a,bpaper ↗); the released files store physical wall names. The mapping between them is set per session, and conflating the two is a frequent source of mislabeling.

Session-normalized stimulus roles
Rolestim_idMeaning
circle10Unrewarded reference in the leaf1-vs-circle1 comparison
circle21Other exemplar, unrewarded family
leaf12Reward-associated texture in task mice; matched but unrewarded role in exposure mice
leaf23Same visual family as leaf1; introduced without reward (later leaf3 = 4, leaf1_swap = 5, 6)
Resolving roles per session

Four physical families exist (leaf, circle, rock, brick). The released WallName takes values rock1/rock2/wood1/wood2, where wood corresponds to the paper’s “brick”. Which physical wall carries which role is set per session. In rewarded sessions the role can be recovered with get_cat_id(WallName, isRew) (rewarded wall → leaf1). In unrewarded sessions isRew is uniformly false, so the role is read directly from stim_id: UniqWalls[stim_id==2] is leaf1 and UniqWalls[stim_id==0] is circle1. Either way, stim_ID=[2,0] denotes leaf1 versus circle1 for every cohort and physical pair.

04
The data

Contents of the release, and the tractable access path

The Nature Data availability statement deposits the study on Figshare; the workspace pins the updated release (doi:10.25378/janelia.28811129.v2, CC BY 4.0), and the paper recipes are in the paper-tagged analysis release. Files are keyed by recording_id = {mouse}_{date}_{block}. The reduced representation reconstructs spk = UTV and supports efficient exploratory population analysis. Exact paper-style per-neuron selectivity uses the full deconvolved traces; agreement between representations must be measured rather than assumed.

Released data layers
LayerFileContentsSize
Imaging behaviourBeh_{experiment}.npy23 experiment bundles linked to imaging acquisitions; session dictionaries contain the aligned frame/trial fields documented below108.1–412.5 MiB
Behaviour-only learningBeh_no_pretrain.npy, Beh_pretrain_on_grat_image.npy, Beh_pretrain_on_nat_image.npy3 bundles for the 23 additional mice reported in Figure 5; no neural or retinotopy files378.7–817.9 MiB
Reduced neural{rec}_SVD_dec.npyU (comp×neuron), V (comp×frame), 400 components59.6–174.9 MiB
Full neural{rec}_neural_data.npy{'spks': [...]} → concat = neurons×frames1.4–9.3 GiB
Retinotopy{rec}_trans.npziarea (neuron→area), xy_t (cortical x,y)0.7–3.1 MiB
Area outlinesareas.npz['out'] area outline polygons158 KB
IndexImaging_Exp_info.npy142 source metadata rows; 133 unique experiment–recording pairs21 KB
Imaging & neurons

The acquisition used a two-photon random-access mesoscope and temporal multiplexing (Methods: imaging acquisition ↗). The first Results section reports simultaneous V1/HVA recordings containing 20,547–89,577 traces (Results: recording scale ↗; Figure 1c–e ↗). Suite2p processing and non-negative deconvolution with a 0.75 s decay parameter are specified in Methods: processing of calcium imaging data ↗. The 400-component SVD files are a deposited-release representation, not a paper Methods claim (Figshare v2 release ↗).

Behaviour fields (per frame)

Released-field inventory: ft_trInd trial, ft_WallID texture, ft_move running, ft_CorrSpc in-texture, ft_Pos/ft_PosCum position, and ft_RunSpeed; per trial, WallName, UniqWalls, stim_id, isRew, ntrials, and SoundDelPos. These field names come from the deposited arrays and the checksum-pinned release audit, not from the article prose (Figshare v2 release ↗; release atlas ↓). The paper defines the corresponding behavioral protocol in Methods: behavioural training ↗.

Visual-area identifiers (neu_area_ID)

Paper-code mapping: V1 = iarea==8; mHV (medial) = {0,1,2,9}, grouping PM, AM, MMA, and lateral retrosplenial cortex; lHV (lateral) = {5,6}; aHV (anterior) = {3,4}; identifiers -1 and 7 are excluded from those groups (paper code: neu_area_ID lines 312–324 ↗). The regional before/after comparison reports the clearest familiar-selective fraction increase in medial HVAs, while the whole-V1 change is smaller (Results: regional plasticity ↗; Figure 1i–j ↗).

05
Availability audit

How much neural data exists, and whether it is day by day

The imaging release described by the Nature article and its Data availability statement contains 89 physical recordings from 19 mice. The workspace pins the updated Figshare v2 release: 297 files and 452,233,500,962 bytes (421.175 GiB). The values below are computed from that immutable inventory rather than inferred from folder names.

The publication originally cited Figshare v1. Version 2 retains those data and adds 89 reduced SVD files plus one neural example; it adds a tractable representation, not 89 new recording sessions.

Neural coverage

89/89 indexed recordings have a full deconvolved-activity file, a 400-component SVD file, and a retinotopy file. These are processed neural traces used by the publication, not raw two-photon movies (Methods: processing of calcium imaging data ↗).

Longitudinal coverage

Not continuous daily imaging. Mice have 1–8 sampled recording dates at protocol landmarks. Trial order is available inside each session; sparse before/after comparisons are available across sessions (Fig. 1bpaper ↗).

Verified Figshare v2 neural layers — release source: Janelia dataset
LayerFilesExact bytesGiBInterpretation
Full deconvolved activity89434,174,046,325404.356Paper-compatible single-neuron response distributions; 1.4–9.3 GiB per recording
Reduced neural (400-PC SVD)8910,746,406,39810.008Fast exploratory reconstruction; 59.6–174.9 MiB per recording
Retinotopy / area assignment89177,062,3020.165Area and cortical-position mapping for every recording
1,000-neuron example derivative197,192,1280.091Small demonstration subset, not another recording
All neural-related files268445,194,707,153414.62098.44% of the complete release by bytes
Do not double-count representations

The full and SVD files are two representations of the same 89 recordings, not 178 independent neural observations. The publication reports 20,547–89,577 neurons per recording and bases its neural analyses on deconvolved fluorescence traces (Methods: calcium processing ↗).

What “day by day” means in this release

There are 89 mouse–date recordings on 77 calendar dates. Eighteen of 19 mice were imaged more than once, with a median of five recording dates per mouse. Across the 70 within-mouse intervals, the median gap is three days (range 1–67); only 21 intervals are one day apart. Those adjacent dates cluster around test-stage boundaries and do not form a complete daily training series. This sampling follows the milestone design shown in Nature Fig. 1bpaper ↗.

The defensible longitudinal claims

Yes: trial-resolved neural change inside a sampled session; before/after population-distribution change paired by mouse. No: a continuous neural learning curve for every training day. Not supplied: an identity map registering the same neuron across dates. Retinotopy aligns recordings to cortical areas, not individual cells across sessions; before/after must therefore be described as mouse/session-paired, never neuron-paired.

Exact Train 1 before/after coverage for the candidate study
Thirteen matched mice and 26 recordings; stages follow the Nature Fig. 1 timelinepaper ↗
CohortMouseBeforeAfterGap
SupervisedTX1082023-03-132023-03-229 d
TX1092023-03-272023-04-0711 d
TX602021-04-102021-05-0424 d
VR22021-03-202021-04-0617 d
UnrewardedDR102022-07-122022-07-197 d
DR152022-10-092022-10-1910 d
TX1042022-10-122022-10-208 d
TX1052022-10-082022-10-1911 d
TX1192023-12-142023-12-239 d
TX1232023-12-212024-01-0212 d
TX832022-08-172022-08-2912 d
TX852022-06-142022-06-173 d
TX882022-06-132022-06-174 d

The four supervised mice and nine unrewarded mice are the same animals before and after Train 1. The dates are sparse endpoints, separated by 3–24 days. The task protocol and the paper’s endpoint result are shown in Nature Fig. 1paper ↗.

All 19 imaging mice: sampled dates and within-mouse gaps
Physical recording coverage from the released index; gaps are ordered days between sampled dates
MouseDatesFirst → lastSuccessive gaps (days)Neural GiB
DR1062022-07-12 → 07-307, 2, 7, 1, 130.687
DR1552022-10-09 → 11-0310, 1, 13, 132.021
LZ1342024-05-15 → 05-291, 12, 111.803
LZ1642024-05-18 → 06-011, 12, 110.756
TX10422022-10-12 → 10-2085.911
TX10552022-10-08 → 11-0811, 2, 15, 326.932
TX10872023-01-05 → 04-0767, 9, 3, 7, 3, 342.082
TX10962023-03-16 → 05-1311, 11, 11, 24, 137.110
TX11982023-12-12 → 2024-01-091, 1, 9, 1, 13, 1, 219.408
TX12382023-12-18 → 2024-01-172, 1, 12, 1, 12, 1, 126.176
TX12432023-12-23 → 12-261, 24.894
TX13922024-05-18 → 05-31136.517
TX14012024-05-31—4.753
TX6052021-04-10 → 06-2224, 34, 14, 120.458
TX6152021-06-07 → 06-2512, 1, 3, 231.624
TX8332022-08-17 → 08-3112, 214.854
TX8522022-06-14 → 06-17312.052
TX8862022-06-13 → 07-224, 3, 25, 4, 335.611
VR272021-03-20 → 05-0617, 5, 17, 1, 5, 240.881

Neural GiB is full activity + SVD + retinotopy. It is a download-size audit, not an experimental sample-size column.

Download footprint for the 13-mouse candidate analysis

Exact preflight for the 26 Train 1 recordings in the four requested strata
PlanIncluded layersDownloadUse
Exploratory26 SVD + 26 retinotopy + 4 shared behaviour bundles3.910 GiBFast graph and pipeline development
Paper-compatible26 full neural + 26 retinotopy + 4 shared behaviour bundles123.038 GiBPrimary single-neuron distribution analysis
Both representationsAll of the above126.080 GiBSVD-to-full validation plus final analysis
06
Complete release atlas

Every experiment, mouse, acquisition, membership, and file

The paper reports 89 recordings in 19 imaging mice (Methods: animals ↗) and deposits the files on Figshare (Data availability ↗). This static atlas is a separate release audit generated from the checksum-pinned Figshare v2 inventory and the deposited Imaging_Exp_info.npy ↗. Its 89 physical acquisitions, 142 source metadata rows, and 23 experiment labels are inventory-derived counts, not numbers copied from a paper paragraph.

File-complete, with one declared manifest limit

The deposited manifest provides every file, acquisition key, exact byte size, MD5, and direct URL, but not each array’s neuron × frame shape. Those shapes must be read from selected SVD/full files after checksum verification (U.shape[1], V.shape[1], or concatenated spks.shape) and are not inferred from file size here. The paper reports 20,547–89,577 traces per recording in the first Results section (Results: supervised and unsupervised plasticity ↗).

Three grains—do not count them as one table

A physical acquisition is identified by recording_id; full and SVD files live at that grain. Retinotopy is acquisition-level but keyed without the block suffix. Behavior is stored once per experiment bundle and contains one or more session keys. The 142 metadata rows reduce to 133 unique experiment–recording memberships and 89 acquisitions; 25 acquisitions are reused under more than one experiment label.

All 23 experiment labels

Association rows, unique physical recordings, mice, exact behavior bundle, and every recording ID
ExperimentSource rowsRecordingsMiceBehavior bundleRecording IDs
naive_test111119Beh_naive_test1.npy
432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3		LZ13_2024_05_15_1 · LZ13_2024_05_16_2 · LZ16_2024_05_18_2 · LZ16_2024_05_19_1 · TX108_2023_01_05_2 · TX109_2023_03_16_1 · TX119_2023_12_12_1 · TX123_2023_12_18_1 · TX124_2023_12_23_1 · TX139_2024_05_18_1 · TX140_2024_05_31_1
naive_test2777Beh_naive_test2.npy
286,584,765 B · 273.309 MiB · MD5 b622db42667fb01b433eef841dd58599		LZ13_2024_05_15_1 · LZ16_2024_05_18_2 · TX119_2023_12_12_1 · TX123_2023_12_18_1 · TX124_2023_12_23_1 · TX139_2024_05_18_1 · TX140_2024_05_31_1
naive_test31053Beh_naive_test3.npy
430,048,721 B · 410.126 MiB · MD5 6261f3c6f0da26630522ca6d0724e4a9		TX119_2023_12_13_1 · TX123_2023_12_18_1 · TX123_2023_12_20_1 · TX124_2023_12_24_1 · TX124_2023_12_26_1
sup_test1555Beh_sup_test1.npy
180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c		TX108_2023_03_25_1 · TX109_2023_04_18_1 · TX60_2021_06_07_1 · TX61_2021_06_07_2 · VR2_2021_04_11_1
sup_test2555Beh_sup_test2.npy
161,695,939 B · 154.205 MiB · MD5 67d5413083b3ffbb5f0b5ceddfd1c66b		TX108_2023_04_04_1 · TX109_2023_05_13_1 · TX60_2021_06_22_1 · TX61_2021_06_20_1 · VR2_2021_04_29_1
sup_test3553Beh_sup_test3.npy
169,187,762 B · 161.350 MiB · MD5 c9f3060140564dc56518c6f56a4683bb		TX108_2023_04_07_1 · TX61_2021_06_23_1 · TX61_2021_06_25_1 · VR2_2021_05_04_1 · VR2_2021_05_06_1
sup_train1_after_learning444Beh_sup_train1_after_learning.npy
113,352,887 B · 108.102 MiB · MD5 08c7066c45c4edfb871ff29b314ba64a		TX108_2023_03_22_1 · TX109_2023_04_07_1 · TX60_2021_05_04_1 · VR2_2021_04_06_1
sup_train1_before_learning444Beh_sup_train1_before_learning.npy
124,559,852 B · 118.790 MiB · MD5 75169b8c4c02f5ed9af3fd492e93b9bd		TX108_2023_03_13_1 · TX109_2023_03_27_1 · TX60_2021_04_10_1 · VR2_2021_03_20_1
sup_train2_after_learning555Beh_sup_train2_after_learning.npy
141,228,534 B · 134.686 MiB · MD5 a1e4bcb573c7b1646daf8830648f4ef6		TX108_2023_04_01_1 · TX109_2023_05_12_1 · TX60_2021_06_21_1 · TX61_2021_06_19_1 · VR2_2021_04_28_1
sup_train2_before_learning555Beh_sup_train2_before_learning.npy
180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c		TX108_2023_03_25_1 · TX109_2023_04_18_1 · TX60_2021_06_07_1 · TX61_2021_06_07_2 · VR2_2021_04_11_1
test1_after_grating553Beh_test1_after_grating.npy
201,376,774 B · 192.048 MiB · MD5 fab95e365509b57b94cd485440fe55f2		LZ13_2024_05_28_1 · LZ13_2024_05_29_1 · LZ16_2024_05_31_2 · LZ16_2024_06_01_1 · TX139_2024_05_31_1
test1_before_grating553Beh_test1_before_grating.npy
178,959,600 B · 170.669 MiB · MD5 aeca11d92b8d294a0011d3beb2190243		LZ13_2024_05_15_1 · LZ13_2024_05_16_2 · LZ16_2024_05_18_2 · LZ16_2024_05_19_1 · TX139_2024_05_18_1
test2_after_grating553Beh_test2_after_grating.npy
201,376,772 B · 192.048 MiB · MD5 a079a211852cd4bca56033feeb54c3a3		LZ13_2024_05_28_1 · LZ13_2024_05_29_1 · LZ16_2024_05_31_2 · LZ16_2024_06_01_1 · TX139_2024_05_31_1
test2_before_grating553Beh_test2_before_grating.npy
178,959,598 B · 170.669 MiB · MD5 8fc0bb4d971e8ad51b72537c821feeb7		LZ13_2024_05_15_1 · LZ13_2024_05_16_2 · LZ16_2024_05_18_2 · LZ16_2024_05_19_1 · TX139_2024_05_18_1
train1_after_grating553Beh_train1_after_grating.npy
201,145,291 B · 191.827 MiB · MD5 9184dddef39aaef10c3259c1e0748a33		LZ13_2024_05_28_1 · LZ13_2024_05_29_1 · LZ16_2024_05_31_2 · LZ16_2024_06_01_1 · TX139_2024_05_31_1
train1_before_grating553Beh_train1_before_grating.npy
178,753,524 B · 170.473 MiB · MD5 a16441ad6ca6af74bb2a504f13dc5003		LZ13_2024_05_15_1 · LZ13_2024_05_16_2 · LZ16_2024_05_18_2 · LZ16_2024_05_19_1 · TX139_2024_05_18_1
unsup_test1777Beh_unsup_test1.npy
270,497,670 B · 257.967 MiB · MD5 00ddb9a93e9972186ebef2521e2568dd		DR10_2022_07_21_1 · DR15_2022_10_20_1 · TX105_2022_10_21_1 · TX119_2023_12_24_1 · TX123_2024_01_03_1 · TX83_2022_08_31_1 · TX88_2022_06_20_1
unsup_test2666Beh_unsup_test2.npy
248,282,556 B · 236.781 MiB · MD5 557f97ba6ec10ec6ab829d4b1314daaa		DR10_2022_07_29_1 · DR15_2022_11_03_2 · TX105_2022_11_08_1 · TX119_2024_01_07_1 · TX123_2024_01_16_1 · TX88_2022_07_19_1
unsup_test3844Beh_unsup_test3.npy
334,494,274 B · 318.999 MiB · MD5 984045b2b31b34326894bb4c2780e1d3		DR10_2022_07_30_1 · TX119_2024_01_09_1 · TX123_2024_01_17_1 · TX88_2022_07_22_1
unsup_train1_after_learning999Beh_unsup_train1_after_learning.npy
329,822,100 B · 314.543 MiB · MD5 470b06d441c6f274a9400a92a3e8e5de		DR10_2022_07_19_1 · DR15_2022_10_19_1 · TX104_2022_10_20_1 · TX105_2022_10_19_2 · TX119_2023_12_23_1 · TX123_2024_01_02_1 · TX83_2022_08_29_1 · TX85_2022_06_17_1 · TX88_2022_06_17_2
unsup_train1_before_learning999Beh_unsup_train1_before_learning.npy
308,224,980 B · 293.946 MiB · MD5 a24dd14d900228ca37f6c9cdbe49e762		DR10_2022_07_12_1 · DR15_2022_10_09_1 · TX104_2022_10_12_1 · TX105_2022_10_08_2 · TX119_2023_12_14_1 · TX123_2023_12_21_1 · TX83_2022_08_17_1 · TX85_2022_06_14_1 · TX88_2022_06_13_2
unsup_train2_after_learning666Beh_unsup_train2_after_learning.npy
225,961,099 B · 215.493 MiB · MD5 b4e7a7d42b366e050903f9137455dd1e		DR10_2022_07_28_1 · DR15_2022_11_02_1 · TX105_2022_11_05_1 · TX119_2024_01_06_1 · TX123_2024_01_15_1 · TX88_2022_07_15_1
unsup_train2_before_learning666Beh_unsup_train2_before_learning.npy
219,700,905 B · 209.523 MiB · MD5 61a001e678b43f0432b97e9a600b3735		DR10_2022_07_21_1 · DR15_2022_10_20_1 · TX105_2022_10_21_1 · TX119_2023_12_24_1 · TX123_2024_01_03_1 · TX88_2022_06_20_1

All 19 mice and 89 neural acquisitions

Every mouse panel is expanded by default. Each acquisition lists its full processed deconvolved activity, 400-component SVD representation, retinotopy transform, linked behavior bundle(s), and every source metadata row. Missing fields remain missing; source spellings and reward-label capitalization are not silently normalized.

DR10 · 6 acquisitions · sex not released · 6 dates (2022-07-12, 2022-07-19, 2022-07-21, 2022-07-28, 2022-07-29, 2022-07-30)
Complete acquisition and logical-membership record for DR10
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
DR10_2022_07_12_12022-07-12 · block 1
unsup_train1_before_learning · behavior key DR10_2022_07_12_1
mname="DR10"datexp="2022_07_12"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0,2]depth=[250,100]is2p=1isDR=1Note="VR move when non-run"ROIdir=[]
DR10_2022_07_12_1_neural_data.npy
7,384,433,945 B · 6.877 GiB · MD5 f632e105fccc0b629f2be62aced34347		DR10_2022_07_12_1_SVD_dec.npy
143,895,262 B · 137.229 MiB · MD5 461787c8e0ac2e8700f5f5b4669d2418		DR10_2022_07_12_trans.npz
2,330,690 B · 2.223 MiB · MD5 44890105820924b884693bdbd206e4b7		Beh_unsup_train1_before_learning.npy308,224,980 B · 293.946 MiB · MD5 a24dd14d900228ca37f6c9cdbe49e762
DR10_2022_07_19_12022-07-19 · block 1
unsup_train1_after_learning · behavior key DR10_2022_07_19_1
mname="DR10"datexp="2022_07_19"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2]depth=[250,100]is2p=1isDR=1Note="close loop VR"ROIdir=[]
DR10_2022_07_19_1_neural_data.npy
4,889,028,665 B · 4.553 GiB · MD5 b2251ad3c1c0af2bad2b853cfd5aaf98		DR10_2022_07_19_1_SVD_dec.npy
123,316,062 B · 117.603 MiB · MD5 4d4f1d71f08a8b108155f3a6a587bd99		DR10_2022_07_19_trans.npz
2,191,370 B · 2.090 MiB · MD5 d1d43cfc545f908997d85cc23b4272bb		Beh_unsup_train1_after_learning.npy329,822,100 B · 314.543 MiB · MD5 470b06d441c6f274a9400a92a3e8e5de
DR10_2022_07_21_12022-07-21 · block 1
unsup_test1 · behavior key DR10_2022_07_21_1
mname="DR10"datexp="2022_07_21"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3]depth=[250,100]is2p=1isDR=1ROIdir=[]
unsup_train2_before_learning · behavior key DR10_2022_07_21_1
mname="DR10"datexp="2022_07_21"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3]depth=[250,100]is2p=1isDR=1ROIdir=[]
DR10_2022_07_21_1_neural_data.npy
4,847,006,857 B · 4.514 GiB · MD5 9eb2e1b5cef6413973e53f5ed9cea9ec		DR10_2022_07_21_1_SVD_dec.npy
119,069,662 B · 113.554 MiB · MD5 a22d2a1e96600a4232cfde56068288f1		DR10_2022_07_21_trans.npz
2,015,610 B · 1.922 MiB · MD5 16c734f5b081d48e481efc6dadb4621d		Beh_unsup_test1.npy270,497,670 B · 257.967 MiB · MD5 00ddb9a93e9972186ebef2521e2568dd
Beh_unsup_train2_before_learning.npy219,700,905 B · 209.523 MiB · MD5 61a001e678b43f0432b97e9a600b3735
DR10_2022_07_28_12022-07-28 · block 1
unsup_train2_after_learning · behavior key DR10_2022_07_28_1
mname="DR10"datexp="2022_07_28"blk="1"exptype="unsup"rewType="None"stim_id=[0,2,3]depth=[250,100]days=6is2p=1isDR=1ROIdir=[]
DR10_2022_07_28_1_neural_data.npy
4,053,347,113 B · 3.775 GiB · MD5 7266a78136e29fd7cd779e08035cd873		DR10_2022_07_28_1_SVD_dec.npy
112,618,462 B · 107.401 MiB · MD5 7a46304be80b7309a8472c99b522326a		DR10_2022_07_28_trans.npz
2,009,530 B · 1.916 MiB · MD5 6e004852b590f8f7fe3205f4efcf11b7		Beh_unsup_train2_after_learning.npy225,961,099 B · 215.493 MiB · MD5 b4e7a7d42b366e050903f9137455dd1e
DR10_2022_07_29_12022-07-29 · block 1
unsup_test2 · behavior key DR10_2022_07_29_1
mname="DR10"datexp="2022_07_29"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2,3,4]depth=[250,100]is2p=1isDR=1ROIdir=[]
DR10_2022_07_29_1_neural_data.npy
6,355,912,577 B · 5.919 GiB · MD5 86ec6f263e2135dffcb7693a046b9741		DR10_2022_07_29_1_SVD_dec.npy
135,088,862 B · 128.831 MiB · MD5 e325f7803c5ea6e6577e3e0d72b1e94a		DR10_2022_07_29_trans.npz
2,245,970 B · 2.142 MiB · MD5 9e1f9ec41a8922b176d3bbdd50423b58		Beh_unsup_test2.npy248,282,556 B · 236.781 MiB · MD5 557f97ba6ec10ec6ab829d4b1314daaa
DR10_2022_07_30_12022-07-30 · block 1
unsup_test3 · behavior key DR10_2022_07_30_1_swap1
mname="DR10"datexp="2022_07_30"blk="1"sess#=1stimtype="swap1"exptype="unsup"rewType="None"stim_id=[0.0,2.0,5.0,null,3.0]depth=[250,100]is2p=1ROIdir=[]
unsup_test3 · behavior key DR10_2022_07_30_1_swap2
mname="DR10"datexp="2022_07_30"blk="1"sess#=1stimtype="swap2"exptype="unsup"rewType="None"stim_id=[0.0,2.0,null,6.0,3.0]depth=[250,100]is2p=1ROIdir=[]
DR10_2022_07_30_1_neural_data.npy
4,657,399,625 B · 4.338 GiB · MD5 4b2c0aa570feee19c6db5a70aaa6a57b		DR10_2022_07_30_1_SVD_dec.npy
115,588,062 B · 110.233 MiB · MD5 bb8e516f77f6cd87f50f89c80c049d9f		DR10_2022_07_30_trans.npz
1,920,250 B · 1.831 MiB · MD5 bd68aba7264645d043ebaef49568ae30		Beh_unsup_test3.npy334,494,274 B · 318.999 MiB · MD5 984045b2b31b34326894bb4c2780e1d3
DR15 · 5 acquisitions · sex not released · 5 dates (2022-10-09, 2022-10-19, 2022-10-20, 2022-11-02, 2022-11-03)
Complete acquisition and logical-membership record for DR15
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
DR15_2022_10_09_12022-10-09 · block 1
unsup_train1_before_learning · behavior key DR15_2022_10_09_1
mname="DR15"datexp="2022_10_09"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0,2]depth=[240,100]is2p=1isDR=1Note="close loop VR"ROIdir=[]
DR15_2022_10_09_1_neural_data.npy
6,550,155,473 B · 6.100 GiB · MD5 baeb8a2bde277cf328b97af4f4ce7fe8		DR15_2022_10_09_1_SVD_dec.npy
136,199,262 B · 129.890 MiB · MD5 18c3b24e10b0439bfd0eb03547ed20e0		DR15_2022_10_09_trans.npz
2,231,570 B · 2.128 MiB · MD5 a6dd81e4397f1b3f4d8e6ae4a74791c4		Beh_unsup_train1_before_learning.npy308,224,980 B · 293.946 MiB · MD5 a24dd14d900228ca37f6c9cdbe49e762
DR15_2022_10_19_12022-10-19 · block 1
unsup_train1_after_learning · behavior key DR15_2022_10_19_1
mname="DR15"datexp="2022_10_19"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2]depth=[240,100]is2p=1isDR=1Note="close loop VR"ROIdir=[]
DR15_2022_10_19_1_neural_data.npy
6,800,051,033 B · 6.333 GiB · MD5 69ecb04a0fd58381a34a9c8c4c58129d		DR15_2022_10_19_1_SVD_dec.npy
140,450,462 B · 133.944 MiB · MD5 a1bc2552a81def33e12d40b741deb7bd		DR15_2022_10_19_trans.npz
2,358,930 B · 2.250 MiB · MD5 b11bf789b022e697169acbad57f7307b		Beh_unsup_train1_after_learning.npy329,822,100 B · 314.543 MiB · MD5 470b06d441c6f274a9400a92a3e8e5de
DR15_2022_10_20_12022-10-20 · block 1
unsup_test1 · behavior key DR15_2022_10_20_1
mname="DR15"datexp="2022_10_20"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3]depth=[240,100]is2p=1isDR=1ROIdir=[]
unsup_train2_before_learning · behavior key DR15_2022_10_20_1
mname="DR15"datexp="2022_10_20"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3]depth=[240,100]is2p=1isDR=1ROIdir=[]
DR15_2022_10_20_1_neural_data.npy
6,884,227,713 B · 6.411 GiB · MD5 3ae4110f8fc3e5eebda18667d7c8cde0		DR15_2022_10_20_1_SVD_dec.npy
145,552,862 B · 138.810 MiB · MD5 e7b94ac934a40185e0679818f26d6b17		DR15_2022_10_20_trans.npz
2,566,930 B · 2.448 MiB · MD5 aa5e91faed1c2f5fb580fd702754b9e5		Beh_unsup_test1.npy270,497,670 B · 257.967 MiB · MD5 00ddb9a93e9972186ebef2521e2568dd
Beh_unsup_train2_before_learning.npy219,700,905 B · 209.523 MiB · MD5 61a001e678b43f0432b97e9a600b3735
DR15_2022_11_02_12022-11-02 · block 1
unsup_train2_after_learning · behavior key DR15_2022_11_02_1
mname="DR15"datexp="2022_11_02"blk="1"exptype="unsup"rewType="None"stim_id=[0,2,3]depth=[240,100]days=9is2p=1isDR=1ROIdir=[]
DR15_2022_11_02_1_neural_data.npy
6,713,997,833 B · 6.253 GiB · MD5 c3c60b04ffef894bb8a1e6570a9474ed		DR15_2022_11_02_1_SVD_dec.npy
138,743,262 B · 132.316 MiB · MD5 24336f562409b9e11c159ccc8801ed11		DR15_2022_11_02_trans.npz
2,303,330 B · 2.197 MiB · MD5 0c6a9f86635cfcf4358eed6d8a25592c		Beh_unsup_train2_after_learning.npy225,961,099 B · 215.493 MiB · MD5 b4e7a7d42b366e050903f9137455dd1e
DR15_2022_11_03_22022-11-03 · block 2
unsup_test2 · behavior key DR15_2022_11_03_2
mname="DR15"datexp="2022_11_03"blk="2"sess#=1exptype="unsup"rewType="None"stim_id=[0,2,3,4]depth=[240,100]is2p=1isDR=1ROIdir=[]
DR15_2022_11_03_2_neural_data.npy
6,720,722,901 B · 6.259 GiB · MD5 66e50dc0def94764c40a5cf190a7665b		DR15_2022_11_03_2_SVD_dec.npy
140,178,462 B · 133.685 MiB · MD5 e6baa27db56133fe8c1ce03c18bb9bd6		DR15_2022_11_03_trans.npz
2,371,770 B · 2.262 MiB · MD5 a73b0ea7f2af80876e13d6c6ada9c634		Beh_unsup_test2.npy248,282,556 B · 236.781 MiB · MD5 557f97ba6ec10ec6ab829d4b1314daaa
LZ13 · 4 acquisitions · Female · 4 dates (2024-05-15, 2024-05-16, 2024-05-28, 2024-05-29)
Complete acquisition and logical-membership record for LZ13
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
LZ13_2024_05_15_12024-05-15 · block 1
naive_test1 · behavior key LZ13_2024_05_15_1
Gender="Female"mname="LZ13"datexp="2024_05_15"blk="1"sess#=1rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
naive_test2 · behavior key LZ13_2024_05_15_1
Gender="Female"mname="LZ13"datexp="2024_05_15"blk="1"sess#=1rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
test1_before_grating · behavior key LZ13_2024_05_15_1
Gender="Female"mname="LZ13"datexp="2024_05_15"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
test2_before_grating · behavior key LZ13_2024_05_15_1
Gender="Female"mname="LZ13"datexp="2024_05_15"blk="1"sess#=0rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
train1_before_grating · behavior key LZ13_2024_05_15_1
Gender="Female"mname="LZ13"datexp="2024_05_15"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0.0,null,2.0,null,null]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
LZ13_2024_05_15_1_neural_data.npy
2,395,977,329 B · 2.231 GiB · MD5 dce0cf4333eda0947c679a36080a6808		LZ13_2024_05_15_1_SVD_dec.npy
83,111,262 B · 79.261 MiB · MD5 45fdf974b9511432f925b50dd925a211		LZ13_2024_05_15_trans.npz
1,248,930 B · 1.191 MiB · MD5 79e270d642c2561d5d70798cd5687c01		Beh_naive_test1.npy432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3
Beh_naive_test2.npy286,584,765 B · 273.309 MiB · MD5 b622db42667fb01b433eef841dd58599
Beh_test1_before_grating.npy178,959,600 B · 170.669 MiB · MD5 aeca11d92b8d294a0011d3beb2190243
Beh_test2_before_grating.npy178,959,598 B · 170.669 MiB · MD5 8fc0bb4d971e8ad51b72537c821feeb7
Beh_train1_before_grating.npy178,753,524 B · 170.473 MiB · MD5 a16441ad6ca6af74bb2a504f13dc5003
LZ13_2024_05_16_22024-05-16 · block 2
naive_test1 · behavior key LZ13_2024_05_16_2
Gender="Female"mname="LZ13"datexp="2024_05_16"blk="2"sess#=2rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
test1_before_grating · behavior key LZ13_2024_05_16_2
Gender="Female"mname="LZ13"datexp="2024_05_16"blk="2"sess#=0exptype="unsup"rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
test2_before_grating · behavior key LZ13_2024_05_16_2
Gender="Female"mname="LZ13"datexp="2024_05_16"blk="2"sess#=0rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
train1_before_grating · behavior key LZ13_2024_05_16_2
Gender="Female"mname="LZ13"datexp="2024_05_16"blk="2"sess#=0exptype="unsup"rewType="None"stim_id=[0.0,null,2.0,null,null]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
LZ13_2024_05_16_2_neural_data.npy
2,745,532,073 B · 2.557 GiB · MD5 26ee99c2d2de60dba9bfb936716a9702		LZ13_2024_05_16_2_SVD_dec.npy
87,088,862 B · 83.054 MiB · MD5 0c7fc3761b384106683e584607b4c996		LZ13_2024_05_16_trans.npz
1,245,942 B · 1.188 MiB · MD5 409634e869f545dc89e59c5e0f5d3cba		Beh_naive_test1.npy432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3
Beh_test1_before_grating.npy178,959,600 B · 170.669 MiB · MD5 aeca11d92b8d294a0011d3beb2190243
Beh_test2_before_grating.npy178,959,598 B · 170.669 MiB · MD5 8fc0bb4d971e8ad51b72537c821feeb7
Beh_train1_before_grating.npy178,753,524 B · 170.473 MiB · MD5 a16441ad6ca6af74bb2a504f13dc5003
LZ13_2024_05_28_12024-05-28 · block 1
test1_after_grating · behavior key LZ13_2024_05_28_1
Gender="Female"mname="LZ13"datexp="2024_05_28"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
test2_after_grating · behavior key LZ13_2024_05_28_1
Gender="Female"mname="LZ13"datexp="2024_05_28"blk="1"sess#=1rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
train1_after_grating · behavior key LZ13_2024_05_28_1
Gender="Female"mname="LZ13"datexp="2024_05_28"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0.0,null,2.0,null,null]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
LZ13_2024_05_28_1_neural_data.npy
3,634,131,853 B · 3.385 GiB · MD5 063d432bc120e06a2cec5a6812e29821		LZ13_2024_05_28_1_SVD_dec.npy
99,356,062 B · 94.753 MiB · MD5 0e2c5c2683530b23434e914e18465759		LZ13_2024_05_28_trans.npz
1,386,846 B · 1.323 MiB · MD5 892eb0d547734ae45ed86585752ff20b		Beh_test1_after_grating.npy201,376,774 B · 192.048 MiB · MD5 fab95e365509b57b94cd485440fe55f2
Beh_test2_after_grating.npy201,376,772 B · 192.048 MiB · MD5 a079a211852cd4bca56033feeb54c3a3
Beh_train1_after_grating.npy201,145,291 B · 191.827 MiB · MD5 9184dddef39aaef10c3259c1e0748a33
LZ13_2024_05_29_12024-05-29 · block 1
test1_after_grating · behavior key LZ13_2024_05_29_1
Gender="Female"mname="LZ13"datexp="2024_05_29"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
test2_after_grating · behavior key LZ13_2024_05_29_1
Gender="Female"mname="LZ13"datexp="2024_05_29"blk="1"sess#=1rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
train1_after_grating · behavior key LZ13_2024_05_29_1
Gender="Female"mname="LZ13"datexp="2024_05_29"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0.0,null,2.0,null,null]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
LZ13_2024_05_29_1_neural_data.npy
3,523,800,929 B · 3.282 GiB · MD5 ce931a9a46be042eeeb06920e1fb9f21		LZ13_2024_05_29_1_SVD_dec.npy
98,791,262 B · 94.215 MiB · MD5 b86a3dc9fd36eec9c9e1022a9d7f97b5		LZ13_2024_05_29_trans.npz
1,418,130 B · 1.352 MiB · MD5 dc0a96d85383c3fc56a380aa1581a68b		Beh_test1_after_grating.npy201,376,774 B · 192.048 MiB · MD5 fab95e365509b57b94cd485440fe55f2
Beh_test2_after_grating.npy201,376,772 B · 192.048 MiB · MD5 a079a211852cd4bca56033feeb54c3a3
Beh_train1_after_grating.npy201,145,291 B · 191.827 MiB · MD5 9184dddef39aaef10c3259c1e0748a33
LZ16 · 4 acquisitions · Female · 4 dates (2024-05-18, 2024-05-19, 2024-05-31, 2024-06-01)
Complete acquisition and logical-membership record for LZ16
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
LZ16_2024_05_18_22024-05-18 · block 2
naive_test1 · behavior key LZ16_2024_05_18_2
Gender="Female"mname="LZ16"datexp="2024_05_18"blk="2"sess#=1rewType="None"stim_id=[0.0,1.0,null,2.0,3.0,null]stim="cirlce0,circle1,circle2,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
naive_test2 · behavior key LZ16_2024_05_18_2
Gender="Female"mname="LZ16"datexp="2024_05_18"blk="2"sess#=1rewType="None"stim_id=[0.0,null,null,2.0,3.0,4.0]stim="cirlce0,circle1,circle2,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
test1_before_grating · behavior key LZ16_2024_05_18_2
Gender="Female"mname="LZ16"datexp="2024_05_18"blk="2"sess#=0exptype="unsup"rewType="None"stim_id=[0.0,1.0,null,2.0,3.0,null]stim="cirlce0,circle1,circle2,leaf0,leaf1,leaf2"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
test2_before_grating · behavior key LZ16_2024_05_18_2
Gender="Female"mname="LZ16"datexp="2024_05_18"blk="2"sess#=0rewType="None"stim_id=[0.0,null,null,2.0,3.0,4.0]stim="cirlce0,circle1,circle2,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
train1_before_grating · behavior key LZ16_2024_05_18_2
Gender="Female"mname="LZ16"datexp="2024_05_18"blk="2"sess#=0exptype="unsup"rewType="None"stim_id=[0.0,null,null,2.0,null,null]stim="cirlce0,circle1,circle2,leaf0,leaf1,leaf2"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
LZ16_2024_05_18_2_neural_data.npy
3,241,121,129 B · 3.019 GiB · MD5 d10a727219f297be6cdbb01fc9d90d72		LZ16_2024_05_18_2_SVD_dec.npy
92,340,062 B · 88.062 MiB · MD5 1093966384ea95883f580e2f1a4f63e5		LZ16_2024_05_18_trans.npz
1,210,014 B · 1.154 MiB · MD5 9b5c1ede899779c504951b742894fd96		Beh_naive_test1.npy432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3
Beh_naive_test2.npy286,584,765 B · 273.309 MiB · MD5 b622db42667fb01b433eef841dd58599
Beh_test1_before_grating.npy178,959,600 B · 170.669 MiB · MD5 aeca11d92b8d294a0011d3beb2190243
Beh_test2_before_grating.npy178,959,598 B · 170.669 MiB · MD5 8fc0bb4d971e8ad51b72537c821feeb7
Beh_train1_before_grating.npy178,753,524 B · 170.473 MiB · MD5 a16441ad6ca6af74bb2a504f13dc5003
LZ16_2024_05_19_12024-05-19 · block 1
naive_test1 · behavior key LZ16_2024_05_19_1
Gender="Female"mname="LZ16"datexp="2024_05_19"blk="1"sess#=2rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
test1_before_grating · behavior key LZ16_2024_05_19_1
Gender="Female"mname="LZ16"datexp="2024_05_19"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
test2_before_grating · behavior key LZ16_2024_05_19_1
Gender="Female"mname="LZ16"datexp="2024_05_19"blk="1"sess#=0rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
train1_before_grating · behavior key LZ16_2024_05_19_1
Gender="Female"mname="LZ16"datexp="2024_05_19"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0.0,null,2.0,null,null]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
LZ16_2024_05_19_1_neural_data.npy
2,191,517,365 B · 2.041 GiB · MD5 958e6f047e4c7e912164187526dcbeba		LZ16_2024_05_19_1_SVD_dec.npy
78,495,262 B · 74.859 MiB · MD5 c1ba297d8fc2f881e51f9044edb2f971		LZ16_2024_05_19_trans.npz
1,148,166 B · 1.095 MiB · MD5 9da3143a7df9d2a4b9a9fd49796c21eb		Beh_naive_test1.npy432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3
Beh_test1_before_grating.npy178,959,600 B · 170.669 MiB · MD5 aeca11d92b8d294a0011d3beb2190243
Beh_test2_before_grating.npy178,959,598 B · 170.669 MiB · MD5 8fc0bb4d971e8ad51b72537c821feeb7
Beh_train1_before_grating.npy178,753,524 B · 170.473 MiB · MD5 a16441ad6ca6af74bb2a504f13dc5003
LZ16_2024_05_31_22024-05-31 · block 2
test1_after_grating · behavior key LZ16_2024_05_31_2
Gender="Female"mname="LZ16"datexp="2024_05_31"blk="2"sess#=1exptype="unsup"rewType="None"stim_id=[0.0,1.0,null,2.0,3.0,null]stim="cirlce0,circle1,circle2,leaf0,leaf1,leaf2"depth=[240,200]is2p=12pblk=["1",2]isDR=0Note="VR move when run"ROIdir=[]
test2_after_grating · behavior key LZ16_2024_05_31_2
Gender="Female"mname="LZ16"datexp="2024_05_31"blk="2"sess#=1rewType="None"stim_id=[0.0,null,null,2.0,3.0,4.0]stim="cirlce0,circle1,circle2,leaf0,leaf1,leaf2"depth=[240,200]is2p=12pblk=["1",2]Note="VR move when run"ROIdir=[]
train1_after_grating · behavior key LZ16_2024_05_31_2
Gender="Female"mname="LZ16"datexp="2024_05_31"blk="2"sess#=1exptype="unsup"rewType="None"stim_id=[0.0,null,null,2.0,null,null]stim="cirlce0,circle1,circle2,leaf0,leaf1,leaf2"depth=[240,200]is2p=12pblk=["1",2]isDR=0Note="VR move when run"ROIdir=[]
LZ16_2024_05_31_2_neural_data.npy
2,937,230,841 B · 2.736 GiB · MD5 e547b8a4b353c637c9120b9a880318f2		LZ16_2024_05_31_2_SVD_dec.npy
90,184,862 B · 86.007 MiB · MD5 9e1aedd8bf8eaff7123921d9c81d8d74		LZ16_2024_05_31_trans.npz
1,294,290 B · 1.234 MiB · MD5 4bd2791035f1996ce335beeeea56d126		Beh_test1_after_grating.npy201,376,774 B · 192.048 MiB · MD5 fab95e365509b57b94cd485440fe55f2
Beh_test2_after_grating.npy201,376,772 B · 192.048 MiB · MD5 a079a211852cd4bca56033feeb54c3a3
Beh_train1_after_grating.npy201,145,291 B · 191.827 MiB · MD5 9184dddef39aaef10c3259c1e0748a33
LZ16_2024_06_01_12024-06-01 · block 1
test1_after_grating · behavior key LZ16_2024_06_01_1
Gender="Female"mname="LZ16"datexp="2024_06_01"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
test2_after_grating · behavior key LZ16_2024_06_01_1
Gender="Female"mname="LZ16"datexp="2024_06_01"blk="1"sess#=1rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
train1_after_grating · behavior key LZ16_2024_06_01_1
Gender="Female"mname="LZ16"datexp="2024_06_01"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0.0,null,2.0,null,null]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
LZ16_2024_06_01_1_neural_data.npy
2,824,692,113 B · 2.631 GiB · MD5 e718fa53064b8017ec9143e419096130		LZ16_2024_06_01_1_SVD_dec.npy
88,487,262 B · 84.388 MiB · MD5 d6d321b4b7200273d962006ddbb36e8c		LZ16_2024_06_01_trans.npz
1,271,718 B · 1.213 MiB · MD5 e8933d9414b6cf37a1040ebd0a93786f		Beh_test1_after_grating.npy201,376,774 B · 192.048 MiB · MD5 fab95e365509b57b94cd485440fe55f2
Beh_test2_after_grating.npy201,376,772 B · 192.048 MiB · MD5 a079a211852cd4bca56033feeb54c3a3
Beh_train1_after_grating.npy201,145,291 B · 191.827 MiB · MD5 9184dddef39aaef10c3259c1e0748a33
TX104 · 2 acquisitions · Male · 2 dates (2022-10-12, 2022-10-20)
Complete acquisition and logical-membership record for TX104
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX104_2022_10_12_12022-10-12 · block 1
unsup_train1_before_learning · behavior key TX104_2022_10_12_1
Gender="Male"mname="TX104"datexp="2022_10_12"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0,2]depth=[240,100]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX104_2022_10_12_1_neural_data.npy
3,395,557,449 B · 3.162 GiB · MD5 08dfbcfc2a041892affd194441fcdda4		TX104_2022_10_12_1_SVD_dec.npy
101,125,662 B · 96.441 MiB · MD5 3f3dcd3ada0097c0b08caf1e04eb2039		TX104_2022_10_12_trans.npz
1,755,090 B · 1.674 MiB · MD5 65aa9b110aa36bf274a8b670d68d73db		Beh_unsup_train1_before_learning.npy308,224,980 B · 293.946 MiB · MD5 a24dd14d900228ca37f6c9cdbe49e762
TX104_2022_10_20_12022-10-20 · block 1
unsup_train1_after_learning · behavior key TX104_2022_10_20_1
Gender="Male"mname="TX104"datexp="2022_10_20"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2]depth=[240,100]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX104_2022_10_20_1_neural_data.npy
2,753,274,933 B · 2.564 GiB · MD5 d7ba105ef03dc8b70b64f72dda91caae		TX104_2022_10_20_1_SVD_dec.npy
93,983,262 B · 89.629 MiB · MD5 d663bb84025de72e82b671e788f62100		TX104_2022_10_20_trans.npz
1,704,330 B · 1.625 MiB · MD5 0da850b7b42f1c39dee6b1366350d317		Beh_unsup_train1_after_learning.npy329,822,100 B · 314.543 MiB · MD5 470b06d441c6f274a9400a92a3e8e5de
TX105 · 5 acquisitions · Male · 5 dates (2022-10-08, 2022-10-19, 2022-10-21, 2022-11-05, 2022-11-08)
Complete acquisition and logical-membership record for TX105
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX105_2022_10_08_22022-10-08 · block 2
unsup_train1_before_learning · behavior key TX105_2022_10_08_2
Gender="Male"mname="TX105"datexp="2022_10_08"blk="2"sess#=0exptype="unsup"rewType="None"stim_id=[0,2]depth=[240,100]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX105_2022_10_08_2_neural_data.npy
6,540,093,833 B · 6.091 GiB · MD5 a2ab5df90678f6716231ecd09fd22ac4		TX105_2022_10_08_2_SVD_dec.npy
138,695,262 B · 132.270 MiB · MD5 436bb45d119981e6b88b246d2be3a2cd		TX105_2022_10_08_trans.npz
2,359,370 B · 2.250 MiB · MD5 925cbea0c97f621a7a9e04c2f652801a		Beh_unsup_train1_before_learning.npy308,224,980 B · 293.946 MiB · MD5 a24dd14d900228ca37f6c9cdbe49e762
TX105_2022_10_19_22022-10-19 · block 2
unsup_train1_after_learning · behavior key TX105_2022_10_19_2
Gender="Male"mname="TX105"datexp="2022_10_19"blk="2"sess#=1exptype="unsup"rewType="None"stim_id=[0,2]depth=[240,100]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX105_2022_10_19_2_neural_data.npy
3,462,284,909 B · 3.225 GiB · MD5 6b4f225aa724efdd0bccb48bcb8b6485		TX105_2022_10_19_2_SVD_dec.npy
113,935,262 B · 108.657 MiB · MD5 787889df3dc14070a3199f391983dc54		TX105_2022_10_19_trans.npz
2,227,850 B · 2.125 MiB · MD5 19556628be34c0413f3a69f0c2876883		Beh_unsup_train1_after_learning.npy329,822,100 B · 314.543 MiB · MD5 470b06d441c6f274a9400a92a3e8e5de
TX105_2022_10_21_12022-10-21 · block 1
unsup_test1 · behavior key TX105_2022_10_21_1
Gender="Male"mname="TX105"datexp="2022_10_21"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3]depth=[240,100]is2p=1isDR=0ROIdir=[]
unsup_train2_before_learning · behavior key TX105_2022_10_21_1
Gender="Male"mname="TX105"datexp="2022_10_21"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3]depth=[240,100]is2p=1isDR=0ROIdir=[]
TX105_2022_10_21_1_neural_data.npy
6,064,808,613 B · 5.648 GiB · MD5 a5b295188ffd57614daf1c5fbdd1d8e4		TX105_2022_10_21_1_SVD_dec.npy
137,484,062 B · 131.115 MiB · MD5 4d8a8ebb7ce499e9e0173abae742c4aa		TX105_2022_10_21_trans.npz
2,446,330 B · 2.333 MiB · MD5 4175d98d2586ca6459a3450dc2e46045		Beh_unsup_test1.npy270,497,670 B · 257.967 MiB · MD5 00ddb9a93e9972186ebef2521e2568dd
Beh_unsup_train2_before_learning.npy219,700,905 B · 209.523 MiB · MD5 61a001e678b43f0432b97e9a600b3735
TX105_2022_11_05_12022-11-05 · block 1
unsup_train2_after_learning · behavior key TX105_2022_11_05_1
Gender="Male"mname="TX105"datexp="2022_11_05"blk="1"exptype="unsup"rewType="None"stim_id=[0,2,3]depth=[240,100]days=13is2p=1isDR=0ROIdir=[]
TX105_2022_11_05_1_neural_data.npy
6,189,381,257 B · 5.764 GiB · MD5 7975f915ffca6f29c84523e0f067dafd		TX105_2022_11_05_1_SVD_dec.npy
127,101,662 B · 121.214 MiB · MD5 2e3936ef643ad5fe93569346ce74c629		TX105_2022_11_05_trans.npz
1,810,010 B · 1.726 MiB · MD5 a9c5385ae48524ad93e301a6f24d1671		Beh_unsup_train2_after_learning.npy225,961,099 B · 215.493 MiB · MD5 b4e7a7d42b366e050903f9137455dd1e
TX105_2022_11_08_12022-11-08 · block 1
unsup_test2 · behavior key TX105_2022_11_08_1
Gender="Male"mname="TX105"datexp="2022_11_08"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2,3,4]depth=[240,100]is2p=1isDR=0ROIdir=[]
TX105_2022_11_08_1_neural_data.npy
6,001,557,013 B · 5.589 GiB · MD5 ffb402f578772ac595a0655aa5efeffc		TX105_2022_11_08_1_SVD_dec.npy
132,024,862 B · 125.909 MiB · MD5 c94e558182b82ff64f5c05c14055b248		TX105_2022_11_08_trans.npz
2,220,010 B · 2.117 MiB · MD5 62c607b32b63fdd0465caef5868dfcaf		Beh_unsup_test2.npy248,282,556 B · 236.781 MiB · MD5 557f97ba6ec10ec6ab829d4b1314daaa
TX108 · 7 acquisitions · Male · 7 dates (2023-01-05, 2023-03-13, 2023-03-22, 2023-03-25, 2023-04-01, 2023-04-04, 2023-04-07)
Complete acquisition and logical-membership record for TX108
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX108_2023_01_05_22023-01-05 · block 2
naive_test1 · behavior key TX108_2023_01_05_2
Gender="Male"mname="TX108"datexp="2023_01_05"blk="2"sess#=1rewType="None"stim_id=[0,1,2,3]stim="cirlce0,circle1,leaf0,leaf1"depth=[250,210]is2p=1Note="VR move when run"ROIdir=[]
TX108_2023_01_05_2_neural_data.npy
6,359,799,389 B · 5.923 GiB · MD5 189783922d03e65f0f60683a742f6b92		TX108_2023_01_05_2_SVD_dec.npy
146,799,266 B · 139.999 MiB · MD5 4c50ab3ca0984ee6024c9d06736351d6		TX108_2023_01_05_trans.npz
2,469,150 B · 2.355 MiB · MD5 475df8c485a8a3dc63fcffbfb583d922		Beh_naive_test1.npy432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3
TX108_2023_03_13_12023-03-13 · block 1
sup_train1_before_learning · behavior key TX108_2023_03_13_1
Gender="Male"mname="TX108"datexp="2023_03_13"blk="1"sess#=0exptype="sup"rewType="Passive"stim_id=[0,2]is2p=1isDR=0ROIdir=[]
TX108_2023_03_13_1_neural_data.npy
7,861,517,081 B · 7.322 GiB · MD5 7198e6b915ab46c6d3e3ffd69603853a		TX108_2023_03_13_1_SVD_dec.npy
173,562,466 B · 165.522 MiB · MD5 66e74de2fa04b44473cb022fc218f369		TX108_2023_03_13_trans.npz
3,078,558 B · 2.936 MiB · MD5 ad11366cca099003be409687efb6a74c		Beh_sup_train1_before_learning.npy124,559,852 B · 118.790 MiB · MD5 75169b8c4c02f5ed9af3fd492e93b9bd
TX108_2023_03_22_12023-03-22 · block 1
sup_train1_after_learning · behavior key TX108_2023_03_22_1
Gender="Male"mname="TX108"datexp="2023_03_22"blk="1"sess#=1exptype="sup"rewType="active after cue"stim_id=[0,2]is2p=1isDR=0ROIdir=[]
TX108_2023_03_22_1_neural_data.npy
5,907,367,193 B · 5.502 GiB · MD5 4cde6c5d3f33c3a1f5c65150c722a242		TX108_2023_03_22_1_SVD_dec.npy
152,079,266 B · 145.034 MiB · MD5 0a933f53cbe7702b89467c66ebc07b34		TX108_2023_03_22_trans.npz
2,718,522 B · 2.593 MiB · MD5 5083ea99f53082ef53df7fb0a1334a50		Beh_sup_train1_after_learning.npy113,352,887 B · 108.102 MiB · MD5 08c7066c45c4edfb871ff29b314ba64a
TX108_2023_03_25_12023-03-25 · block 1
sup_test1 · behavior key TX108_2023_03_25_1
Gender="Male"mname="TX108"datexp="2023_03_25"blk="1"sess#=1exptype="sup"rewType="active after cue"stim_id=[0,1,2,3]depth=[200,240]is2p=1isDR=0ROIdir=[]
sup_train2_before_learning · behavior key TX108_2023_03_25_1
Gender="Male"mname="TX108"datexp="2023_03_25"blk="1"sess#=1exptype="sup"rewType="active after cue"stim_id=[0,1,2,3]depth=[200,240]is2p=1isDR=0ROIdir=[]
TX108_2023_03_25_1_neural_data.npy
8,449,969,273 B · 7.870 GiB · MD5 3e87d3fcc74b2f841e4645cc78d0eac1		TX108_2023_03_25_1_SVD_dec.npy
169,632,866 B · 161.775 MiB · MD5 29900e2e23ffada97c2c5d8fecfe8858		TX108_2023_03_25_trans.npz
2,860,254 B · 2.728 MiB · MD5 3f5d70c150aba8adf84166ae8a419dbb		Beh_sup_test1.npy180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c
Beh_sup_train2_before_learning.npy180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c
TX108_2023_04_01_12023-04-01 · block 1
sup_train2_after_learning · behavior key TX108_2023_04_01_1
Gender="Male"mname="TX108"datexp="2023_04_01"blk="1"sess#=6exptype="sup"rewType="active after cue"stim_id=[0,2,3]is2p=1isDR=0ROIdir=[]
TX108_2023_04_01_1_neural_data.npy
4,662,716,053 B · 4.342 GiB · MD5 e2d13aff87ae14001a5c4b51dfa2687d		TX108_2023_04_01_1_SVD_dec.npy
136,287,266 B · 129.974 MiB · MD5 a6bf999b328f61d27963818b72b2b470		TX108_2023_04_01_trans.npz
2,450,862 B · 2.337 MiB · MD5 acb6e3e8995a6881830b9bbed73bd0f2		Beh_sup_train2_after_learning.npy141,228,534 B · 134.686 MiB · MD5 a1e4bcb573c7b1646daf8830648f4ef6
TX108_2023_04_04_12023-04-04 · block 1
sup_test2 · behavior key TX108_2023_04_04_1
Gender="Male"mname="TX108"datexp="2023_04_04"blk="1"sess#=1exptype="sup"rewType="Active after cue"stim_id=[0,2,3,4]depth=[250,210]is2p=1isDR=0ROIdir=[]
TX108_2023_04_04_1_neural_data.npy
6,205,358,993 B · 5.779 GiB · MD5 a18234e9ca549404edb590d84aa26f83		TX108_2023_04_04_1_SVD_dec.npy
154,887,266 B · 147.712 MiB · MD5 ef498a3a52858adcba7b1136baebce03		TX108_2023_04_04_trans.npz
2,756,322 B · 2.629 MiB · MD5 dafe16f94a57770f1645ebd804626eba		Beh_sup_test2.npy161,695,939 B · 154.205 MiB · MD5 67d5413083b3ffbb5f0b5ceddfd1c66b
TX108_2023_04_07_12023-04-07 · block 1
sup_test3 · behavior key TX108_2023_04_07_1_swap2
Gender="Male"mname="TX108"datexp="2023_04_07"blk="1"sess#=1stimtype="swap2"exptype="sup"rewType="Active after cue"stim_id=[0,2,6,3]is2p=1ROIdir=[]
TX108_2023_04_07_1_neural_data.npy
4,652,948,401 B · 4.333 GiB · MD5 281d7e64e39b4b8dab7c6fba17178825		TX108_2023_04_07_1_SVD_dec.npy
133,642,466 B · 127.451 MiB · MD5 5e979c934a54ff0a62f4a90657758029		TX108_2023_04_07_trans.npz
2,372,238 B · 2.262 MiB · MD5 3f63205e0dac401e54ee522ecac8d180		Beh_sup_test3.npy169,187,762 B · 161.350 MiB · MD5 c9f3060140564dc56518c6f56a4683bb
TX109 · 6 acquisitions · Male · 6 dates (2023-03-16, 2023-03-27, 2023-04-07, 2023-04-18, 2023-05-12, 2023-05-13)
Complete acquisition and logical-membership record for TX109
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX109_2023_03_16_12023-03-16 · block 1
naive_test1 · behavior key TX109_2023_03_16_1
Gender="Male"mname="TX109"datexp="2023_03_16"blk="1"sess#=1rewType="None"stim_id=[2,3,0,1]stim="cirlce0,circle1,leaf0,leaf1"depth=[250,210]is2p=1Note="VR move when run"ROIdir=[]
TX109_2023_03_16_1_neural_data.npy
9,970,697,353 B · 9.286 GiB · MD5 04665840e42c799c268acf4c75f838bd		TX109_2023_03_16_1_SVD_dec.npy
183,418,466 B · 174.921 MiB · MD5 8aefb4ef058c783120bdc0fb317cb78d		TX109_2023_03_16_trans.npz
3,078,090 B · 2.935 MiB · MD5 84ee4896005ee1096095baaf0397916f		Beh_naive_test1.npy432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3
TX109_2023_03_27_12023-03-27 · block 1
sup_train1_before_learning · behavior key TX109_2023_03_27_1
Gender="Male"mname="TX109"datexp="2023_03_27"blk="1"sess#=0exptype="sup"rewType="Passive"stim_id=[2,0]is2p=1isDR=0ROIdir=[]
TX109_2023_03_27_1_neural_data.npy
3,712,137,473 B · 3.457 GiB · MD5 d8fd40b04b883aade63818031236ac65		TX109_2023_03_27_1_SVD_dec.npy
125,143,262 B · 119.346 MiB · MD5 d2191482783b70b27a1d4a6cc79ad507		TX109_2023_03_27_trans.npz
2,291,742 B · 2.186 MiB · MD5 d5808278e74325de72f6ff67b5d93c38		Beh_sup_train1_before_learning.npy124,559,852 B · 118.790 MiB · MD5 75169b8c4c02f5ed9af3fd492e93b9bd
TX109_2023_04_07_12023-04-07 · block 1
sup_train1_after_learning · behavior key TX109_2023_04_07_1
Gender="Male"mname="TX109"datexp="2023_04_07"blk="1"sess#=1exptype="sup"rewType="active after cue"stim_id=[2,0]is2p=1isDR=0ROIdir=[]
TX109_2023_04_07_1_neural_data.npy
7,173,684,941 B · 6.681 GiB · MD5 b6a44ef7ff410d8d711808afa533815c		TX109_2023_04_07_1_SVD_dec.npy
175,362,466 B · 167.239 MiB · MD5 805f141dec3a536abaf0015d837ab049		TX109_2023_04_07_trans.npz
3,226,014 B · 3.077 MiB · MD5 2258c5a2bdd9706ba023fc14cc18a353		Beh_sup_train1_after_learning.npy113,352,887 B · 108.102 MiB · MD5 08c7066c45c4edfb871ff29b314ba64a
TX109_2023_04_18_12023-04-18 · block 1
sup_test1 · behavior key TX109_2023_04_18_1
Gender="Male"mname="TX109"datexp="2023_04_18"blk="1"sess#=1exptype="sup"rewType="Active after cue"stim_id=[2,3,0,1]depth=[250,210]is2p=1isDR=0ROIdir=[]
sup_train2_before_learning · behavior key TX109_2023_04_18_1
Gender="Male"mname="TX109"datexp="2023_04_18"blk="1"sess#=1exptype="sup"rewType="Active after cue"stim_id=[2,3,0,1]depth=[250,210]is2p=1isDR=0ROIdir=[]
TX109_2023_04_18_1_neural_data.npy
5,788,643,029 B · 5.391 GiB · MD5 e5e5130d9f16d446592d9e908a4d73a4		TX109_2023_04_18_1_SVD_dec.npy
161,093,666 B · 153.631 MiB · MD5 101b7bf2307fe0196c04d4c42cbbaca4		TX109_2023_04_18_trans.npz
3,000,366 B · 2.861 MiB · MD5 85fc6357a55d316509f70ac2b0403732		Beh_sup_test1.npy180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c
Beh_sup_train2_before_learning.npy180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c
TX109_2023_05_12_12023-05-12 · block 1
sup_train2_after_learning · behavior key TX109_2023_05_12_1
Gender="Male"mname="TX109"datexp="2023_05_12"blk="1"exptype="sup"rewType="active after cue"stim_id=[2,3,0]days=15is2p=1isDR=0ROIdir=[]
TX109_2023_05_12_1_neural_data.npy
5,440,470,277 B · 5.067 GiB · MD5 562090e8dbe4a9db88fa132de3e3dc52		TX109_2023_05_12_1_SVD_dec.npy
145,061,666 B · 138.342 MiB · MD5 2334386cdb37214b5cd03b2e4d87ce59		TX109_2023_05_12_trans.npz
2,581,974 B · 2.462 MiB · MD5 0349f03cf226456a8267e1cc3554ac4d		Beh_sup_train2_after_learning.npy141,228,534 B · 134.686 MiB · MD5 a1e4bcb573c7b1646daf8830648f4ef6
TX109_2023_05_13_12023-05-13 · block 1
sup_test2 · behavior key TX109_2023_05_13_1
Gender="Male"mname="TX109"datexp="2023_05_13"blk="1"sess#=1exptype="sup"rewType="active after cue"stim_id=[2,3,4,0,1]is2p=1isDR=0ROIdir=[]
TX109_2023_05_13_1_neural_data.npy
6,795,837,785 B · 6.329 GiB · MD5 a929d60496baab9574938ca6770d5ca5		TX109_2023_05_13_1_SVD_dec.npy
157,708,066 B · 150.402 MiB · MD5 a47aa3389fec1abdf1748d00a77bc8b5		TX109_2023_05_13_trans.npz
2,747,898 B · 2.621 MiB · MD5 3e9599de71c35444baedaf72b596ff76		Beh_sup_test2.npy161,695,939 B · 154.205 MiB · MD5 67d5413083b3ffbb5f0b5ceddfd1c66b
TX119 · 8 acquisitions · Male · 8 dates (2023-12-12, 2023-12-13, 2023-12-14, 2023-12-23, 2023-12-24, 2024-01-06, 2024-01-07, 2024-01-09)
Complete acquisition and logical-membership record for TX119
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX119_2023_12_12_12023-12-12 · block 1
naive_test1 · behavior key TX119_2023_12_12_1
Gender="Male"mname="TX119"datexp="2023_12_12"blk="1"sess#=1rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
naive_test2 · behavior key TX119_2023_12_12_1
mname="TX119"datexp="2023_12_12"blk="1"sess#=1rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="rock0,rock1,brick0,brick1,brick5"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
TX119_2023_12_12_1_neural_data.npy
2,094,023,553 B · 1.950 GiB · MD5 8f40c078efd213e5adbc8850bdfe50fd		TX119_2023_12_12_1_SVD_dec.npy
75,232,862 B · 71.748 MiB · MD5 6cbcde0e70350f1e203ae50c016aba9d		TX119_2023_12_12_trans.npz
1,041,894 B · 1,017.475 KiB · MD5 033ef0d3ecc930acc88d84d162b27c72		Beh_naive_test1.npy432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3
Beh_naive_test2.npy286,584,765 B · 273.309 MiB · MD5 b622db42667fb01b433eef841dd58599
TX119_2023_12_13_12023-12-13 · block 1
naive_test3 · behavior key TX119_2023_12_13_1_swap1
Gender="Male"mname="TX119"datexp="2023_12_13"blk="1"sess#=1stimtype="swap1"exptype="naive"rewType="None"stim_id=[0.0,2.0,5.0,null,3.0]depth=[240,200]is2p=1ROIdir=[]
naive_test3 · behavior key TX119_2023_12_13_1_swap2
Gender="Male"mname="TX119"datexp="2023_12_13"blk="1"sess#=2stimtype="swap2"exptype="naive"rewType="None"stim_id=[0.0,2.0,null,6.0,3.0]depth=[240,200]is2p=1ROIdir=[]
TX119_2023_12_13_1_neural_data.npy
3,126,522,393 B · 2.912 GiB · MD5 33a2620ad2c75750c1a91f816eccf3f1		TX119_2023_12_13_1_SVD_dec.npy
91,304,862 B · 87.075 MiB · MD5 715d409c28a1b7f4a3898222be97de7c		TX119_2023_12_13_trans.npz
1,233,234 B · 1.176 MiB · MD5 9e0703f4c4ba35abfd7341ea49a01457		Beh_naive_test3.npy430,048,721 B · 410.126 MiB · MD5 6261f3c6f0da26630522ca6d0724e4a9
TX119_2023_12_14_12023-12-14 · block 1
unsup_train1_before_learning · behavior key TX119_2023_12_14_1
Gender="Male"mname="TX119"datexp="2023_12_14"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0,2]depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX119_2023_12_14_1_neural_data.npy
1,955,049,689 B · 1.821 GiB · MD5 0dd1606a6163bbe7bd4ec5333ccb571a		TX119_2023_12_14_1_SVD_dec.npy
75,372,062 B · 71.880 MiB · MD5 77d2aad59ddbc60c645dbf57136dc0e5		TX119_2023_12_14_trans.npz
1,141,434 B · 1.089 MiB · MD5 d69af9cb080b9e3576c80b9e209a4d5a		Beh_unsup_train1_before_learning.npy308,224,980 B · 293.946 MiB · MD5 a24dd14d900228ca37f6c9cdbe49e762
TX119_2023_12_23_12023-12-23 · block 1
unsup_train1_after_learning · behavior key TX119_2023_12_23_1
Gender="Male"mname="TX119"datexp="2023_12_23"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2]depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX119_2023_12_23_1_neural_data.npy
2,305,926,393 B · 2.148 GiB · MD5 2d302ad0b33acb709682aebedc3cc96a		TX119_2023_12_23_1_SVD_dec.npy
84,074,462 B · 80.180 MiB · MD5 843ad752a61c1940a6c1266ed427ceb7		TX119_2023_12_23_trans.npz
1,330,902 B · 1.269 MiB · MD5 ee770f6195cf3539c86292ae0c4f79d5		Beh_unsup_train1_after_learning.npy329,822,100 B · 314.543 MiB · MD5 470b06d441c6f274a9400a92a3e8e5de
TX119_2023_12_24_12023-12-24 · block 1
unsup_test1 · behavior key TX119_2023_12_24_1
Gender="Male"mname="TX119"datexp="2023_12_24"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3]depth=[240,200]is2p=1isDR=0ROIdir=[]
unsup_train2_before_learning · behavior key TX119_2023_12_24_1
Gender="Male"mname="TX119"datexp="2023_12_24"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3]depth=[240,200]is2p=1isDR=0ROIdir=[]
TX119_2023_12_24_1_neural_data.npy
1,589,937,153 B · 1.481 GiB · MD5 02b9788c79000792c615e711b62d373e		TX119_2023_12_24_1_SVD_dec.npy
66,967,262 B · 63.865 MiB · MD5 633aa6978689a18f8d124a8443c8a8e6		TX119_2023_12_24_trans.npz
983,358 B · 960.311 KiB · MD5 ddda2db80ae338435ffa73b289690ae0		Beh_unsup_test1.npy270,497,670 B · 257.967 MiB · MD5 00ddb9a93e9972186ebef2521e2568dd
Beh_unsup_train2_before_learning.npy219,700,905 B · 209.523 MiB · MD5 61a001e678b43f0432b97e9a600b3735
TX119_2024_01_06_12024-01-06 · block 1
unsup_train2_after_learning · behavior key TX119_2024_01_06_1
Gender="Male"mname="TX119"datexp="2024_01_06"blk="1"sess#=10exptype="unsup"rewType="None"stim_id=[0,2,3]depth=[240,200]is2p=1isDR=0ROIdir=[]
TX119_2024_01_06_1_neural_data.npy
1,967,417,625 B · 1.832 GiB · MD5 91269c3ec16f1dae81bf062c52a2bdc0		TX119_2024_01_06_1_SVD_dec.npy
73,943,262 B · 70.518 MiB · MD5 2c9c7b9d951273df780f4430cc9b37f2		TX119_2024_01_06_trans.npz
1,066,374 B · 1.017 MiB · MD5 469b5449871ccb74ecb80f39594c0ef8		Beh_unsup_train2_after_learning.npy225,961,099 B · 215.493 MiB · MD5 b4e7a7d42b366e050903f9137455dd1e
TX119_2024_01_07_12024-01-07 · block 1
unsup_test2 · behavior key TX119_2024_01_07_1
Gender="Male"mname="TX119"datexp="2024_01_07"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2,3,4]depth=[240,200]is2p=1isDR=0ROIdir=[]
TX119_2024_01_07_1_neural_data.npy
3,100,427,693 B · 2.887 GiB · MD5 292d5d6aa7432092251bc9449678482b		TX119_2024_01_07_1_SVD_dec.npy
90,184,862 B · 86.007 MiB · MD5 1c56361be45b24b7e094fbe530aee5c3		TX119_2024_01_07_trans.npz
1,172,934 B · 1.119 MiB · MD5 1af25f93370527f659f8495238d4aa21		Beh_unsup_test2.npy248,282,556 B · 236.781 MiB · MD5 557f97ba6ec10ec6ab829d4b1314daaa
TX119_2024_01_09_12024-01-09 · block 1
unsup_test3 · behavior key TX119_2024_01_09_1_swap1
Gender="Male"mname="TX119"datexp="2024_01_09"blk="1"sess#=1stimtype="swap1"exptype="unsup"rewType="None"stim_id=[0.0,2.0,5.0,null,3.0]depth=[240,200]is2p=1ROIdir=[]
unsup_test3 · behavior key TX119_2024_01_09_1_swap2
Gender="Male"mname="TX119"datexp="2024_01_09"blk="1"sess#=2stimtype="swap2"exptype="unsup"rewType="None"stim_id=[0.0,2.0,null,6.0,3.0]depth=[240,200]is2p=1ROIdir=[]
TX119_2024_01_09_1_neural_data.npy
4,030,825,689 B · 3.754 GiB · MD5 d5bfc129ee8e0eca470fe079c0c82d6c		TX119_2024_01_09_1_SVD_dec.npy
102,479,262 B · 97.732 MiB · MD5 f56ea2a0df9557d510db447f88cbd6aa		TX119_2024_01_09_trans.npz
1,305,810 B · 1.245 MiB · MD5 529b5aa56b6d28b6df721beb28acf83b		Beh_unsup_test3.npy334,494,274 B · 318.999 MiB · MD5 984045b2b31b34326894bb4c2780e1d3
TX123 · 8 acquisitions · Male · 8 dates (2023-12-18, 2023-12-20, 2023-12-21, 2024-01-02, 2024-01-03, 2024-01-15, 2024-01-16, 2024-01-17)
Complete acquisition and logical-membership record for TX123
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX123_2023_12_18_12023-12-18 · block 1
naive_test1 · behavior key TX123_2023_12_18_1
Gender="Male"mname="TX123"datexp="2023_12_18"blk="1"sess#=1rewType="None"stim_id=[0.0,1.0,null,2.0,null,null,3.0,null]stim="cirlce0,circle1,circle2,leaf0,leaf0_swap1,leaf0_swap2,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
naive_test2 · behavior key TX123_2023_12_18_1
mname="TX123"datexp="2023_12_18"blk="1"sess#=1rewType="None"stim_id=[0.0,null,null,2.0,null,null,3.0,4.0]stim="cirlce0,circle1,circle2,leaf0,leaf0_swap1,leaf0_swap2,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
naive_test3 · behavior key TX123_2023_12_18_1_swap1
Gender="Male"mname="TX123"datexp="2023_12_18"blk="1"sess#=1stimtype="swap1"exptype="naive"rewType="None"stim_id=[0.0,null,null,2.0,5.0,null,3.0,null]stim="cirlce0,circle1,circle2,leaf0,leaf0_swap1,leaf0_swap2,leaf1,leaf2"depth=[240,200]is2p=1ROIdir=[]
naive_test3 · behavior key TX123_2023_12_18_1_swap2
Gender="Male"mname="TX123"datexp="2023_12_18"blk="1"sess#=2stimtype="swap2"exptype="naive"rewType="None"stim_id=[0.0,null,null,2.0,null,6.0,3.0,null]stim="cirlce0,circle1,circle2,leaf0,leaf0_swap1,leaf0_swap2,leaf1,leaf2"depth=[240,200]is2p=1ROIdir=[]
TX123_2023_12_18_1_neural_data.npy
6,989,509,673 B · 6.509 GiB · MD5 352ea57c8a8da1bafb93088b6543204a		TX123_2023_12_18_1_SVD_dec.npy
136,996,062 B · 130.650 MiB · MD5 11b4f6da5b672889b5670cb4076412f9		TX123_2023_12_18_trans.npz
1,874,826 B · 1.788 MiB · MD5 9d8b8906294aebcc7c7a2f2d38e3fa04		Beh_naive_test1.npy432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3
Beh_naive_test2.npy286,584,765 B · 273.309 MiB · MD5 b622db42667fb01b433eef841dd58599
Beh_naive_test3.npy430,048,721 B · 410.126 MiB · MD5 6261f3c6f0da26630522ca6d0724e4a9
TX123_2023_12_20_12023-12-20 · block 1
naive_test3 · behavior key TX123_2023_12_20_1_swap1
mname="TX123"datexp="2023_12_20"blk="1"sess#=3stimtype="swap1"exptype="naive"rewType="None"stim_id=[0.0,2.0,5.0,null,3.0]stim="rock0, brick0, brick1, brick0_swap1, brick0_swap2"depth=[240,200]is2p=1ROIdir=[]
naive_test3 · behavior key TX123_2023_12_20_1_swap2
Gender="Male"mname="TX123"datexp="2023_12_20"blk="1"sess#=4stimtype="swap2"exptype="naive"rewType="None"stim_id=[0.0,2.0,null,6.0,3.0]stim="rock0, brick0, brick1, brick0_swap1, brick0_swap2"depth=[240,200]is2p=1ROIdir=[]
TX123_2023_12_20_1_neural_data.npy
4,690,360,273 B · 4.368 GiB · MD5 00b9344a256fd8ee998c4b8f6f8487e0		TX123_2023_12_20_1_SVD_dec.npy
115,104,862 B · 109.773 MiB · MD5 53baa10874d88acac30adf846867f42b		TX123_2023_12_20_trans.npz
1,692,342 B · 1.614 MiB · MD5 541466b93863df3095598fd082a6fcf8		Beh_naive_test3.npy430,048,721 B · 410.126 MiB · MD5 6261f3c6f0da26630522ca6d0724e4a9
TX123_2023_12_21_12023-12-21 · block 1
unsup_train1_before_learning · behavior key TX123_2023_12_21_1
Gender="Male"mname="TX123"datexp="2023_12_21"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0,2]depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX123_2023_12_21_1_neural_data.npy
1,830,346,137 B · 1.705 GiB · MD5 72fc878d935bdee0b2d362a689168718		TX123_2023_12_21_1_SVD_dec.npy
71,432,862 B · 68.124 MiB · MD5 583d2463381459e92f759f624ffaa9ce		TX123_2023_12_21_trans.npz
1,034,298 B · 1,010.057 KiB · MD5 0a15078f0479b438af0b7f96ac979191		Beh_unsup_train1_before_learning.npy308,224,980 B · 293.946 MiB · MD5 a24dd14d900228ca37f6c9cdbe49e762
TX123_2024_01_02_12024-01-02 · block 1
unsup_train1_after_learning · behavior key TX123_2024_01_02_1
Gender="Male"mname="TX123"datexp="2024_01_02"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2]depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX123_2024_01_02_1_neural_data.npy
5,043,223,885 B · 4.697 GiB · MD5 00083b7e1fddab0aeed35a50559de45a		TX123_2024_01_02_1_SVD_dec.npy
115,772,062 B · 110.409 MiB · MD5 7b73ef8242870c323304bd865f11f447		TX123_2024_01_02_trans.npz
1,552,950 B · 1.481 MiB · MD5 575e35ba218e046895ba1a1c5caa4997		Beh_unsup_train1_after_learning.npy329,822,100 B · 314.543 MiB · MD5 470b06d441c6f274a9400a92a3e8e5de
TX123_2024_01_03_12024-01-03 · block 1
unsup_test1 · behavior key TX123_2024_01_03_1
Gender="Male"mname="TX123"datexp="2024_01_03"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3]depth=[240,200]is2p=1isDR=0ROIdir=[]
unsup_train2_before_learning · behavior key TX123_2024_01_03_1
Gender="Male"mname="TX123"datexp="2024_01_03"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3]depth=[240,200]is2p=1isDR=0ROIdir=[]
TX123_2024_01_03_1_neural_data.npy
2,357,808,741 B · 2.196 GiB · MD5 49cec90bcb560a9c90574d1bc275828e		TX123_2024_01_03_1_SVD_dec.npy
79,986,462 B · 76.281 MiB · MD5 a3410a79550afe048cea3c283e46c25f		TX123_2024_01_03_trans.npz
1,114,758 B · 1.063 MiB · MD5 0b6174b6c021858df87a4840246473ee		Beh_unsup_test1.npy270,497,670 B · 257.967 MiB · MD5 00ddb9a93e9972186ebef2521e2568dd
Beh_unsup_train2_before_learning.npy219,700,905 B · 209.523 MiB · MD5 61a001e678b43f0432b97e9a600b3735
TX123_2024_01_15_12024-01-15 · block 1
unsup_train2_after_learning · behavior key TX123_2024_01_15_1
Gender="Male"mname="TX123"datexp="2024_01_15"blk="1"sess#=12exptype="unsup"rewType="None"stim_id=[0,2,3]depth=[240,200]is2p=1isDR=0ROIdir=[]
TX123_2024_01_15_1_neural_data.npy
1,936,249,481 B · 1.803 GiB · MD5 be6aa4560a9e3d4e34951516b31af26e		TX123_2024_01_15_1_SVD_dec.npy
74,010,462 B · 70.582 MiB · MD5 1c7b2ac454a64d01934927b78c0f6fd4		TX123_2024_01_15_trans.npz
1,090,314 B · 1.040 MiB · MD5 4c837ed7081cac195de85bbd64f401e1		Beh_unsup_train2_after_learning.npy225,961,099 B · 215.493 MiB · MD5 b4e7a7d42b366e050903f9137455dd1e
TX123_2024_01_16_12024-01-16 · block 1
unsup_test2 · behavior key TX123_2024_01_16_1
Gender="Male"mname="TX123"datexp="2024_01_16"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2,3,4]depth=[240,200]is2p=1isDR=0ROIdir=[]
TX123_2024_01_16_1_neural_data.npy
1,844,666,137 B · 1.718 GiB · MD5 142e9bd5990e5bb71d26f3cf9598d53e		TX123_2024_01_16_1_SVD_dec.npy
70,392,862 B · 67.132 MiB · MD5 00d39aee4c9248a46f89385021dc1a9e		TX123_2024_01_16_trans.npz
964,458 B · 941.854 KiB · MD5 b779c82c8db24a82cc4aa8be45be5d8f		Beh_unsup_test2.npy248,282,556 B · 236.781 MiB · MD5 557f97ba6ec10ec6ab829d4b1314daaa
TX123_2024_01_17_12024-01-17 · block 1
unsup_test3 · behavior key TX123_2024_01_17_1_swap1
Gender="Male"mname="TX123"datexp="2024_01_17"blk="1"sess#=1stimtype="swap1"exptype="unsup"rewType="None"stim_id=[0.0,2.0,5.0,null,3.0]depth=[240,200]is2p=1ROIdir=[]
unsup_test3 · behavior key TX123_2024_01_17_1_swap2
Gender="Male"mname="TX123"datexp="2024_01_17"blk="1"sess#=2stimtype="swap2"exptype="unsup"rewType="None"stim_id=[0.0,2.0,null,6.0,3.0]depth=[240,200]is2p=1ROIdir=[]
TX123_2024_01_17_1_neural_data.npy
2,654,857,953 B · 2.473 GiB · MD5 d8901ab564bde48078294eee29044078		TX123_2024_01_17_1_SVD_dec.npy
84,580,062 B · 80.662 MiB · MD5 b9cd14231db2df0ef7842ad7b579b376		TX123_2024_01_17_trans.npz
1,165,086 B · 1.111 MiB · MD5 f93c0745bd886fa0e69af6b12acf9fd4		Beh_unsup_test3.npy334,494,274 B · 318.999 MiB · MD5 984045b2b31b34326894bb4c2780e1d3
TX124 · 3 acquisitions · Male · 3 dates (2023-12-23, 2023-12-24, 2023-12-26)
Complete acquisition and logical-membership record for TX124
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX124_2023_12_23_12023-12-23 · block 1
naive_test1 · behavior key TX124_2023_12_23_1
Gender="Male"mname="TX124"datexp="2023_12_23"blk="1"sess#=1rewType="None"stim_id=[0.0,1.0,null,2.0,3.0,null]stim="cirlce0,circle1,circle2,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
naive_test2 · behavior key TX124_2023_12_23_1
mname="TX124"datexp="2023_12_23"blk="1"sess#=1rewType="None"stim_id=[0.0,null,null,2.0,3.0,4.0]stim="cirlce0,circle1,circle2,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
TX124_2023_12_23_1_neural_data.npy
1,910,749,885 B · 1.780 GiB · MD5 ff5ffca5dc8b1c41f3bd1c8cb4e4f153		TX124_2023_12_23_1_SVD_dec.npy
70,364,062 B · 67.104 MiB · MD5 8413a4248c40f94cfff8e45b84cb6b06		TX124_2023_12_23_trans.npz
879,030 B · 858.428 KiB · MD5 a46a0315fd4ebff958b63cbc0dd3f3d0		Beh_naive_test1.npy432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3
Beh_naive_test2.npy286,584,765 B · 273.309 MiB · MD5 b622db42667fb01b433eef841dd58599
TX124_2023_12_24_12023-12-24 · block 1
naive_test3 · behavior key TX124_2023_12_24_1_swap1
Gender="Male"mname="TX124"datexp="2023_12_24"blk="1"sess#=1stimtype="swap1"exptype="naive"rewType="None"stim_id=[0.0,2.0,5.0,null,3.0]stim="cell0, leaf0, leaf1, leaf0_swap1, leaf0_swap2"depth=[240,200]is2p=1ROIdir=[]
naive_test3 · behavior key TX124_2023_12_24_1_swap2
Gender="Male"mname="TX124"datexp="2023_12_24"blk="1"sess#=2stimtype="swap2"exptype="naive"rewType="None"stim_id=[0.0,2.0,null,6.0,3.0]stim="cell0, leaf0, leaf1, leaf0_swap1, leaf0_swap2"depth=[240,200]is2p=1ROIdir=[]
TX124_2023_12_24_1_neural_data.npy
1,519,738,781 B · 1.415 GiB · MD5 23ce3e823d5eb8c10aac9fce1e29f7a1		TX124_2023_12_24_1_SVD_dec.npy
62,466,462 B · 59.573 MiB · MD5 60fad2f997348362b09f1fc9acf440df		TX124_2023_12_24_trans.npz
740,934 B · 723.568 KiB · MD5 88b2f745771d3a0bccbf359649834a43		Beh_naive_test3.npy430,048,721 B · 410.126 MiB · MD5 6261f3c6f0da26630522ca6d0724e4a9
TX124_2023_12_26_12023-12-26 · block 1
naive_test3 · behavior key TX124_2023_12_26_1_swap1
Gender="Male"mname="TX124"datexp="2023_12_26"blk="1"sess#=1stimtype="swap1"exptype="naive"rewType="None"stim_id=[0.0,2.0,5.0,null,3.0]stim="rock0, brick0, brick1, brick0_swap1, brick0_swap2"depth=[240,200]is2p=1ROIdir=[]
naive_test3 · behavior key TX124_2023_12_26_1_swap2
Gender="Male"mname="TX124"datexp="2023_12_26"blk="1"sess#=2stimtype="swap2"exptype="naive"rewType="None"stim_id=[0.0,2.0,null,6.0,3.0]stim="rock0, brick0, brick1, brick0_swap1, brick0_swap2"depth=[240,200]is2p=1ROIdir=[]
TX124_2023_12_26_1_neural_data.npy
1,624,267,257 B · 1.513 GiB · MD5 ef36aca83ca1c541f1267fd5d10c86f3		TX124_2023_12_26_1_SVD_dec.npy
65,120,862 B · 62.104 MiB · MD5 0ddd173c2ab4e729aaa1eb45c1f932f7		TX124_2023_12_26_trans.npz
835,578 B · 815.994 KiB · MD5 60aa86c3f119d71eeb9ec19250c57b79		Beh_naive_test3.npy430,048,721 B · 410.126 MiB · MD5 6261f3c6f0da26630522ca6d0724e4a9
TX139 · 2 acquisitions · Male · 2 dates (2024-05-18, 2024-05-31)
Complete acquisition and logical-membership record for TX139
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX139_2024_05_18_12024-05-18 · block 1
naive_test1 · behavior key TX139_2024_05_18_1
Gender="Male"mname="TX139"datexp="2024_05_18"blk="1"sess#=1rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
naive_test2 · behavior key TX139_2024_05_18_1
Gender="Male"mname="TX139"datexp="2024_05_18"blk="1"sess#=1rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
test1_before_grating · behavior key TX139_2024_05_18_1
Gender="Male"mname="TX139"datexp="2024_05_18"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
test2_before_grating · behavior key TX139_2024_05_18_1
Gender="Male"mname="TX139"datexp="2024_05_18"blk="1"sess#=0rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
train1_before_grating · behavior key TX139_2024_05_18_1
Gender="Male"mname="TX139"datexp="2024_05_18"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0.0,null,2.0,null,null]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX139_2024_05_18_1_neural_data.npy
3,085,889,825 B · 2.874 GiB · MD5 7f5701a7f3365c2dbdf84a20c2ee3773		TX139_2024_05_18_1_SVD_dec.npy
91,348,062 B · 87.116 MiB · MD5 45bcd57d45c6137be1f79ad5e082b138		TX139_2024_05_18_trans.npz
1,265,778 B · 1.207 MiB · MD5 569c0a73fa20f45c88f7e5a4187697a4		Beh_naive_test1.npy432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3
Beh_naive_test2.npy286,584,765 B · 273.309 MiB · MD5 b622db42667fb01b433eef841dd58599
Beh_test1_before_grating.npy178,959,600 B · 170.669 MiB · MD5 aeca11d92b8d294a0011d3beb2190243
Beh_test2_before_grating.npy178,959,598 B · 170.669 MiB · MD5 8fc0bb4d971e8ad51b72537c821feeb7
Beh_train1_before_grating.npy178,753,524 B · 170.473 MiB · MD5 a16441ad6ca6af74bb2a504f13dc5003
TX139_2024_05_31_12024-05-31 · block 1
test1_after_grating · behavior key TX139_2024_05_31_1
Gender="Male"mname="TX139"datexp="2024_05_31"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
test2_after_grating · behavior key TX139_2024_05_31_1
Gender="Male"mname="TX139"datexp="2024_05_31"blk="1"sess#=1rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
train1_after_grating · behavior key TX139_2024_05_31_1
Gender="Male"mname="TX139"datexp="2024_05_31"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0.0,null,2.0,null,null]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX139_2024_05_31_1_neural_data.npy
3,719,124,577 B · 3.464 GiB · MD5 b3be1c7cd09a8a9a0a147c0982ac5113		TX139_2024_05_31_1_SVD_dec.npy
98,333,662 B · 93.778 MiB · MD5 3ca56bfcb8cdb33eea098bfa67a90009		TX139_2024_05_31_trans.npz
1,244,034 B · 1.186 MiB · MD5 00a2a0b8302cf6491220fd2283e5ca2b		Beh_test1_after_grating.npy201,376,774 B · 192.048 MiB · MD5 fab95e365509b57b94cd485440fe55f2
Beh_test2_after_grating.npy201,376,772 B · 192.048 MiB · MD5 a079a211852cd4bca56033feeb54c3a3
Beh_train1_after_grating.npy201,145,291 B · 191.827 MiB · MD5 9184dddef39aaef10c3259c1e0748a33
TX140 · 1 acquisition · Female · 1 date (2024-05-31)
Complete acquisition and logical-membership record for TX140
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX140_2024_05_31_12024-05-31 · block 1
naive_test1 · behavior key TX140_2024_05_31_1
Gender="Female"mname="TX140"datexp="2024_05_31"blk="1"sess#=1rewType="None"stim_id=[0.0,1.0,2.0,3.0,null]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
naive_test2 · behavior key TX140_2024_05_31_1
Gender="Female"mname="TX140"datexp="2024_05_31"blk="1"sess#=1rewType="None"stim_id=[0.0,null,2.0,3.0,4.0]stim="cirlce0,circle1,leaf0,leaf1,leaf2"depth=[240,200]is2p=1Note="VR move when run"ROIdir=[]
TX140_2024_05_31_1_neural_data.npy
4,982,581,157 B · 4.640 GiB · MD5 267d3e0660b3d6306ecb8846cfe15aa7		TX140_2024_05_31_1_SVD_dec.npy
119,320,862 B · 113.793 MiB · MD5 beef506a82de5ba82d89b9a93a350783		TX140_2024_05_31_trans.npz
1,776,438 B · 1.694 MiB · MD5 b36b8991596a5aab8de0ae3e6afeda72		Beh_naive_test1.npy432,561,586 B · 412.523 MiB · MD5 c582b1937ee1917dce1512036b346ba3
Beh_naive_test2.npy286,584,765 B · 273.309 MiB · MD5 b622db42667fb01b433eef841dd58599
TX60 · 5 acquisitions · Female · 5 dates (2021-04-10, 2021-05-04, 2021-06-07, 2021-06-21, 2021-06-22)
Complete acquisition and logical-membership record for TX60
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX60_2021_04_10_12021-04-10 · block 1
sup_train1_before_learning · behavior key TX60_2021_04_10_1
Gender="Female"mname="TX60"datexp="2021_04_10"blk="1"sess#=0exptype="sup"rewType="Passive"stim_id=[0,2]is2p=1isDR=0artLick=1ROIdir=[]
TX60_2021_04_10_1_neural_data.npy
4,574,765,705 B · 4.261 GiB · MD5 98425630ac650ccbb40c21bd1ff95ab2		TX60_2021_04_10_1_SVD_dec.npy
115,527,262 B · 110.175 MiB · MD5 ab46727703cfb2248017486088fbf794		TX60_2021_04_10_trans.npz
1,755,594 B · 1.674 MiB · MD5 de474ee378baa4d10d3e880c7e3b3cfd		Beh_sup_train1_before_learning.npy124,559,852 B · 118.790 MiB · MD5 75169b8c4c02f5ed9af3fd492e93b9bd
TX60_2021_05_04_12021-05-04 · block 1
sup_train1_after_learning · behavior key TX60_2021_05_04_1
Gender="Female"mname="TX60"datexp="2021_05_04"blk="1"sess#=1exptype="sup"rewType="Passive"stim_id=[0,2]is2p=1isDR=0artLick=1ROIdir=[]
TX60_2021_05_04_1_neural_data.npy
4,033,200,345 B · 3.756 GiB · MD5 24fce049441942e91ae5427cb6c3544e		TX60_2021_05_04_1_SVD_dec.npy
119,916,062 B · 114.361 MiB · MD5 b80f47e80503341f3993a60eb9357f32		TX60_2021_05_04_trans.npz
2,066,490 B · 1.971 MiB · MD5 50d07ddbd9cdd8e2f2fd44da2783ab43		Beh_sup_train1_after_learning.npy113,352,887 B · 108.102 MiB · MD5 08c7066c45c4edfb871ff29b314ba64a
TX60_2021_06_07_12021-06-07 · block 1
sup_test1 · behavior key TX60_2021_06_07_1
Gender="Female"mname="TX60"datexp="2021_06_07"blk="1"sess#=1exptype="sup"rewType="Passive"stim_id=[0,1,2,3]depth=[200,240]is2p=1isDR=0ROIdir=[]
sup_train2_before_learning · behavior key TX60_2021_06_07_1
Gender="Female"mname="TX60"datexp="2021_06_07"blk="1"sess#=1exptype="sup"rewType="Passive"stim_id=[0,1,2,3]depth=[200,240]is2p=1isDR=0ROIdir=[]
TX60_2021_06_07_1_neural_data.npy
4,295,871,973 B · 4.001 GiB · MD5 63a0d2140c8361f47d818f2f7d67d557		TX60_2021_06_07_1_SVD_dec.npy
112,421,662 B · 107.214 MiB · MD5 edbfd59a754af26ce30f79faa7f73823		TX60_2021_06_07_trans.npz
1,913,130 B · 1.825 MiB · MD5 55661f8dcaccd159a4420cf6d0b1cc98		Beh_sup_test1.npy180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c
Beh_sup_train2_before_learning.npy180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c
TX60_2021_06_21_12021-06-21 · block 1
sup_train2_after_learning · behavior key TX60_2021_06_21_1
Gender="Female"mname="TX60"datexp="2021_06_21"blk="1"exptype="sup"rewType="Passive"stim_id=[0,2,3]days=10is2p=1isDR=0ROIdir=[]
TX60_2021_06_21_1_neural_data.npy
4,672,140,185 B · 4.351 GiB · MD5 cdbcd6cc9c7fb780cfc792990e0d22a5		TX60_2021_06_21_1_SVD_dec.npy
128,159,262 B · 122.222 MiB · MD5 b553bb0561257bd3704b3585b0ba7ff1		TX60_2021_06_21_trans.npz
2,194,506 B · 2.093 MiB · MD5 afc99ad622e3c8b4f5ccab2d1654bd55		Beh_sup_train2_after_learning.npy141,228,534 B · 134.686 MiB · MD5 a1e4bcb573c7b1646daf8830648f4ef6
TX60_2021_06_22_12021-06-22 · block 1
sup_test2 · behavior key TX60_2021_06_22_1
Gender="Female"mname="TX60"datexp="2021_06_22"blk="1"sess#=1exptype="sup"rewType="Passive"stim_id=[0,2,3,4]is2p=1isDR=0ROIdir=[]
TX60_2021_06_22_1_neural_data.npy
3,795,442,529 B · 3.535 GiB · MD5 cca22dccdd2e12755fb80a7a4304a1cb		TX60_2021_06_22_1_SVD_dec.npy
109,140,062 B · 104.084 MiB · MD5 4f46177e0266f9ed5bf5b99709fcd3c1		TX60_2021_06_22_trans.npz
1,951,410 B · 1.861 MiB · MD5 dd217c1116c45a5d2f726d70c9c662ea		Beh_sup_test2.npy161,695,939 B · 154.205 MiB · MD5 67d5413083b3ffbb5f0b5ceddfd1c66b
TX61 · 5 acquisitions · Female · 5 dates (2021-06-07, 2021-06-19, 2021-06-20, 2021-06-23, 2021-06-25)
Complete acquisition and logical-membership record for TX61
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX61_2021_06_07_22021-06-07 · block 2
sup_test1 · behavior key TX61_2021_06_07_2
Gender="Female"mname="TX61"datexp="2021_06_07"blk="2"sess#=1exptype="sup"rewType="Passive"stim_id=[0,1,2,3]depth=[200,240]is2p=1isDR=0ROIdir=[]
sup_train2_before_learning · behavior key TX61_2021_06_07_2
Gender="Female"mname="TX61"datexp="2021_06_07"blk="2"sess#=1exptype="sup"rewType="Passive"stim_id=[0,1,2,3]depth=[200,240]is2p=1isDR=0ROIdir=[]
TX61_2021_06_07_2_neural_data.npy
5,860,282,265 B · 5.458 GiB · MD5 11de2b01fd04aac3e5e5aecf0bbe1c01		TX61_2021_06_07_2_SVD_dec.npy
137,268,062 B · 130.909 MiB · MD5 56306c5264d446ab423b38730080169e		TX61_2021_06_07_trans.npz
2,491,970 B · 2.377 MiB · MD5 2714d74d8d28dd855fa4af36fa23da33		Beh_sup_test1.npy180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c
Beh_sup_train2_before_learning.npy180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c
TX61_2021_06_19_12021-06-19 · block 1
sup_train2_after_learning · behavior key TX61_2021_06_19_1
Gender="Female"mname="TX61"datexp="2021_06_19"blk="1"exptype="sup"rewType="Passive"stim_id=[0,2,3]days=8is2p=1isDR=0ROIdir=[]
TX61_2021_06_19_1_neural_data.npy
4,881,371,457 B · 4.546 GiB · MD5 5f20f05d09ce476d4fd7ec85514dc8f3		TX61_2021_06_19_1_SVD_dec.npy
137,130,466 B · 130.778 MiB · MD5 27eaaf07e4d61d65d97ea6465abe0bbe		TX61_2021_06_19_trans.npz
2,437,326 B · 2.324 MiB · MD5 a49bc7bd34ea32a834c451916d3e10ba		Beh_sup_train2_after_learning.npy141,228,534 B · 134.686 MiB · MD5 a1e4bcb573c7b1646daf8830648f4ef6
TX61_2021_06_20_12021-06-20 · block 1
sup_test2 · behavior key TX61_2021_06_20_1
Gender="Female"mname="TX61"datexp="2021_06_20"blk="1"sess#=1exptype="sup"rewType="Passive"stim_id=[0,2,3,4]is2p=1isDR=0ROIdir=[]
TX61_2021_06_20_1_neural_data.npy
5,385,887,045 B · 5.016 GiB · MD5 11f59b8022cae74d48d954cf67797646		TX61_2021_06_20_1_SVD_dec.npy
145,356,066 B · 138.622 MiB · MD5 1a7faff26b391ea6d5c50fd10bb724e5		TX61_2021_06_20_trans.npz
2,889,450 B · 2.756 MiB · MD5 0ec83bf9fb065fa01ecede28126a4b1d		Beh_sup_test2.npy161,695,939 B · 154.205 MiB · MD5 67d5413083b3ffbb5f0b5ceddfd1c66b
TX61_2021_06_23_12021-06-23 · block 1
sup_test3 · behavior key TX61_2021_06_23_1_swap1
Gender="Female"mname="TX61"datexp="2021_06_23"blk="1"sess#=1stimtype="swap1"exptype="sup"rewType="Passive"stim_id=[0,2,5,3]is2p=1ROIdir=[]
TX61_2021_06_23_1_neural_data.npy
7,416,370,033 B · 6.907 GiB · MD5 2a01604cdb1596cfaec4ec280c1f4a1e		TX61_2021_06_23_1_SVD_dec.npy
153,351,266 B · 146.247 MiB · MD5 33e31a4e86f9a620c290f0d67e58362c		TX61_2021_06_23_trans.npz
2,483,478 B · 2.368 MiB · MD5 b30f23bd0d46555b48262adf97e691ec		Beh_sup_test3.npy169,187,762 B · 161.350 MiB · MD5 c9f3060140564dc56518c6f56a4683bb
TX61_2021_06_25_12021-06-25 · block 1
sup_test3 · behavior key TX61_2021_06_25_1_swap2
Gender="Female"mname="TX61"datexp="2021_06_25"blk="1"sess#=1stimtype="swap2"exptype="sup"rewType="Passive"stim_id=[0,2,6,3]is2p=1ROIdir=[]
TX61_2021_06_25_1_neural_data.npy
9,649,962,145 B · 8.987 GiB · MD5 288c161a0ae73af3122e4eb0e75b90d1		TX61_2021_06_25_1_SVD_dec.npy
175,844,066 B · 167.698 MiB · MD5 ae15ff220e5459347d3e4e873634c78d		TX61_2021_06_25_trans.npz
2,866,338 B · 2.734 MiB · MD5 be82ccd8010a2e095cb3649bdee147fa		Beh_sup_test3.npy169,187,762 B · 161.350 MiB · MD5 c9f3060140564dc56518c6f56a4683bb
TX83 · 3 acquisitions · Female · 3 dates (2022-08-17, 2022-08-29, 2022-08-31)
Complete acquisition and logical-membership record for TX83
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX83_2022_08_17_12022-08-17 · block 1
unsup_train1_before_learning · behavior key TX83_2022_08_17_1
Gender="Female"mname="TX83"datexp="2022_08_17"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0,2]stim="VR move when run"depth=[250,100]is2p=1isDR=0ROIdir=[]
TX83_2022_08_17_1_neural_data.npy
3,600,149,481 B · 3.353 GiB · MD5 7253da285ae1cb2b4d135fa28f3843aa		TX83_2022_08_17_1_SVD_dec.npy
103,184,862 B · 98.405 MiB · MD5 b7e97d1440a5386c54cebd8ee0987d55		TX83_2022_08_17_trans.npz
1,764,090 B · 1.682 MiB · MD5 62481e5642a977feab649dfbcc9bd972		Beh_unsup_train1_before_learning.npy308,224,980 B · 293.946 MiB · MD5 a24dd14d900228ca37f6c9cdbe49e762
TX83_2022_08_29_12022-08-29 · block 1
unsup_train1_after_learning · behavior key TX83_2022_08_29_1
Gender="Female"mname="TX83"datexp="2022_08_29"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2]depth=[250,100]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX83_2022_08_29_1_neural_data.npy
5,169,793,585 B · 4.815 GiB · MD5 362095f49108765a6d0ae8580e5f65d5		TX83_2022_08_29_1_SVD_dec.npy
123,783,262 B · 118.049 MiB · MD5 020d8c7f14e7e7040943093907959515		TX83_2022_08_29_trans.npz
2,119,890 B · 2.022 MiB · MD5 872b59d8e5f479814b2f4ed4c2f21770		Beh_unsup_train1_after_learning.npy329,822,100 B · 314.543 MiB · MD5 470b06d441c6f274a9400a92a3e8e5de
TX83_2022_08_31_12022-08-31 · block 1
unsup_test1 · behavior key TX83_2022_08_31_1
Gender="Female"mname="TX83"datexp="2022_08_31"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3]depth=[250,100]is2p=1isDR=0ROIdir=[]
TX83_2022_08_31_1_neural_data.npy
6,811,993,805 B · 6.344 GiB · MD5 f34473c32466c9dee4ebe46988251f39		TX83_2022_08_31_1_SVD_dec.npy
134,863,262 B · 128.616 MiB · MD5 47d23560b2baad68ed716e46d3be2796		TX83_2022_08_31_trans.npz
2,029,290 B · 1.935 MiB · MD5 2097f1df59ab7ac34cc0866833ecd1a4		Beh_unsup_test1.npy270,497,670 B · 257.967 MiB · MD5 00ddb9a93e9972186ebef2521e2568dd
TX85 · 2 acquisitions · Male · 2 dates (2022-06-14, 2022-06-17)
Complete acquisition and logical-membership record for TX85
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX85_2022_06_14_12022-06-14 · block 1
unsup_train1_before_learning · behavior key TX85_2022_06_14_1
Gender="Male"mname="TX85"datexp="2022_06_14"blk="1"sess#=0exptype="unsup"rewType="None"stim_id=[0,2]is2p=1isDR=0ROIdir=[]
TX85_2022_06_14_1_neural_data.npy
6,024,613,593 B · 5.611 GiB · MD5 fab809192c8542ebf44c7a9c5bb9502c		TX85_2022_06_14_1_SVD_dec.npy
135,762,462 B · 129.473 MiB · MD5 0a3ae65f1d439c36a7016576b6b81210		TX85_2022_06_14_trans.npz
2,384,130 B · 2.274 MiB · MD5 534c1ce9ddea1cfaef468903aad376f5		Beh_unsup_train1_before_learning.npy308,224,980 B · 293.946 MiB · MD5 a24dd14d900228ca37f6c9cdbe49e762
TX85_2022_06_17_12022-06-17 · block 1
unsup_train1_after_learning · behavior key TX85_2022_06_17_1
Gender="Male"mname="TX85"datexp="2022_06_17"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2]depth=[250,100]is2p=1isDR=0ROIdir=[]
TX85_2022_06_17_1_neural_data.npy
6,639,639,805 B · 6.184 GiB · MD5 1454cbf9b1f9e38e1aa26173e60f3666		TX85_2022_06_17_1_SVD_dec.npy
136,412,062 B · 130.093 MiB · MD5 e2b8eb87824022401b1238c925fd124e		TX85_2022_06_17_trans.npz
2,208,250 B · 2.106 MiB · MD5 cb08e359878224a2a0db1407b12e173a		Beh_unsup_train1_after_learning.npy329,822,100 B · 314.543 MiB · MD5 470b06d441c6f274a9400a92a3e8e5de
TX88 · 6 acquisitions · Male · 6 dates (2022-06-13, 2022-06-17, 2022-06-20, 2022-07-15, 2022-07-19, 2022-07-22)
Complete acquisition and logical-membership record for TX88
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
TX88_2022_06_13_22022-06-13 · block 2
unsup_train1_before_learning · behavior key TX88_2022_06_13_2
Gender="Male"mname="TX88"datexp="2022_06_13"blk="2"sess#=0exptype="unsup"rewType="None"stim_id=[0,2]depth=[250,100]is2p=12pblk=["2",1]isDR=0Note="VR move when run"ROIdir=[]
TX88_2022_06_13_2_neural_data.npy
5,143,459,009 B · 4.790 GiB · MD5 24cb75b97470f08c87a8ccf5d00ed593		TX88_2022_06_13_2_SVD_dec.npy
133,037,662 B · 126.875 MiB · MD5 f09943c331a3e1444e4ec3710befaf55		TX88_2022_06_13_trans.npz
2,505,970 B · 2.390 MiB · MD5 1d3a4097fd376483a5b110910b395292		Beh_unsup_train1_before_learning.npy308,224,980 B · 293.946 MiB · MD5 a24dd14d900228ca37f6c9cdbe49e762
TX88_2022_06_17_22022-06-17 · block 2
unsup_train1_after_learning · behavior key TX88_2022_06_17_2
Gender="Male"mname="TX88"datexp="2022_06_17"blk="2"sess#=1exptype="unsup"rewType="None"stim_id=[0,2]depth=[250,100]is2p=1isDR=0Note="VR move when run"ROIdir=[]
TX88_2022_06_17_2_neural_data.npy
5,462,234,297 B · 5.087 GiB · MD5 651998020f9badeacf761fa9c3c20591		TX88_2022_06_17_2_SVD_dec.npy
122,874,462 B · 117.182 MiB · MD5 32909867be696eb8fc6d4d00454f68ad		TX88_2022_06_17_trans.npz
1,954,690 B · 1.864 MiB · MD5 99dae06fc8789cc6026867a402741dbd		Beh_unsup_train1_after_learning.npy329,822,100 B · 314.543 MiB · MD5 470b06d441c6f274a9400a92a3e8e5de
TX88_2022_06_20_12022-06-20 · block 1
unsup_test1 · behavior key TX88_2022_06_20_1
Gender="Male"mname="TX88"datexp="2022_06_20"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3,4]depth=[250,100]is2p=1isDR=0ROIdir=[]
unsup_train2_before_learning · behavior key TX88_2022_06_20_1
Gender="Male"mname="TX88"datexp="2022_06_20"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,1,2,3,4]depth=[250,100]is2p=1isDR=0ROIdir=[]
TX88_2022_06_20_1_neural_data.npy
6,837,555,585 B · 6.368 GiB · MD5 d198714a964ed60bb585a9c30a0e9493		TX88_2022_06_20_1_SVD_dec.npy
140,039,262 B · 133.552 MiB · MD5 ace34ee202a09d9cd20eea4d696296c2		TX88_2022_06_20_trans.npz
2,325,690 B · 2.218 MiB · MD5 10c4570f8429491712741e15304cda31		Beh_unsup_test1.npy270,497,670 B · 257.967 MiB · MD5 00ddb9a93e9972186ebef2521e2568dd
Beh_unsup_train2_before_learning.npy219,700,905 B · 209.523 MiB · MD5 61a001e678b43f0432b97e9a600b3735
TX88_2022_07_15_12022-07-15 · block 1
unsup_train2_after_learning · behavior key TX88_2022_07_15_1
Gender="Male"mname="TX88"datexp="2022_07_15"blk="1"exptype="unsup"rewType="None"stim_id=[0,2,3]depth=[250,100]days=7is2p=1isDR=0ROIdir=[]
TX88_2022_07_15_1_neural_data.npy
6,098,008,361 B · 5.679 GiB · MD5 6e1325c4106b9d8c1247c375b29907ef		TX88_2022_07_15_1_SVD_dec.npy
130,848,862 B · 124.787 MiB · MD5 b1deddc23d99b34e27943684d9529bc2		TX88_2022_07_15_trans.npz
2,122,850 B · 2.025 MiB · MD5 d5173145fd6135630ba6bfa56daf8ad0		Beh_unsup_train2_after_learning.npy225,961,099 B · 215.493 MiB · MD5 b4e7a7d42b366e050903f9137455dd1e
TX88_2022_07_19_12022-07-19 · block 1
unsup_test2 · behavior key TX88_2022_07_19_1
Gender="Male"mname="TX88"datexp="2022_07_19"blk="1"sess#=1exptype="unsup"rewType="None"stim_id=[0,2,3,4]depth=[250,100]is2p=1isDR=0ROIdir=[]
TX88_2022_07_19_1_neural_data.npy
7,633,081,497 B · 7.109 GiB · MD5 051481e771f653ed58bb9ccc9d48539b		TX88_2022_07_19_1_SVD_dec.npy
148,016,862 B · 141.160 MiB · MD5 1faba812c457a71153a129252319e301		TX88_2022_07_19_trans.npz
2,459,970 B · 2.346 MiB · MD5 0ee7c138cfb4a6051d95e6e09286b4c9		Beh_unsup_test2.npy248,282,556 B · 236.781 MiB · MD5 557f97ba6ec10ec6ab829d4b1314daaa
TX88_2022_07_22_12022-07-22 · block 1
unsup_test3 · behavior key TX88_2022_07_22_1_swap1
Gender="Male"mname="TX88"datexp="2022_07_22"blk="1"sess#=2stimtype="swap1"exptype="unsup"rewType="None"stim_id=[0.0,2.0,5.0,null,3.0]depth=[250,100]is2p=1ROIdir=[]
unsup_test3 · behavior key TX88_2022_07_22_1_swap2
Gender="Male"mname="TX88"datexp="2022_07_22"blk="1"sess#=2stimtype="swap2"exptype="unsup"rewType="None"stim_id=[0.0,2.0,null,6.0,3.0]depth=[250,100]is2p=1ROIdir=[]
TX88_2022_07_22_1_neural_data.npy
6,240,536,025 B · 5.812 GiB · MD5 0b891435990cef795d76d5d6bee3179d		TX88_2022_07_22_1_SVD_dec.npy
133,855,262 B · 127.654 MiB · MD5 bac1e23fecde1e0df0f4425bdd39a747		TX88_2022_07_22_trans.npz
2,225,410 B · 2.122 MiB · MD5 5485859ceccf01152d8a4e6b5f5317c2		Beh_unsup_test3.npy334,494,274 B · 318.999 MiB · MD5 984045b2b31b34326894bb4c2780e1d3
VR2 · 7 acquisitions · Male · 7 dates (2021-03-20, 2021-04-06, 2021-04-11, 2021-04-28, 2021-04-29, 2021-05-04, 2021-05-06)
Complete acquisition and logical-membership record for VR2
AcquisitionEvery released membership and raw metadataFull neuralSVDRetinotopyBehavior bundle(s)
VR2_2021_03_20_12021-03-20 · block 1
sup_train1_before_learning · behavior key VR2_2021_03_20_1
Gender="Male"mname="VR2"datexp="2021_03_20"blk="1"sess#=0exptype="sup"rewType="Passive"stim_id=[0,2]is2p=1isDR=0artLick=1Note="1-no cue in non-reward corridor"ROIdir=[]
VR2_2021_03_20_1_neural_data.npy
7,918,524,289 B · 7.375 GiB · MD5 318c8dab7b16985038650c712af82e83		VR2_2021_03_20_1_SVD_dec.npy
169,239,266 B · 161.399 MiB · MD5 8ecba1a1d76f4f5a231dcdff38d6e54f		VR2_2021_03_20_trans.npz
2,934,270 B · 2.798 MiB · MD5 f8fbb33ee2c9461011306c5072d0b06e		Beh_sup_train1_before_learning.npy124,559,852 B · 118.790 MiB · MD5 75169b8c4c02f5ed9af3fd492e93b9bd
VR2_2021_04_06_12021-04-06 · block 1
sup_train1_after_learning · behavior key VR2_2021_04_06_1
Gender="Male"mname="VR2"datexp="2021_04_06"blk="1"sess#=1exptype="sup"rewType="Passive"stim_id=[0,2]is2p=1isDR=0artLick=1Note="1-no cue in non-reward corridor"ROIdir=[]
VR2_2021_04_06_1_neural_data.npy
5,048,742,233 B · 4.702 GiB · MD5 b367e8cb78da35fdf59ab4ed87c4eb68		VR2_2021_04_06_1_SVD_dec.npy
142,132,066 B · 135.548 MiB · MD5 9a9554cc07ab5ad34b750ca7ea0e6920		VR2_2021_04_06_trans.npz
2,559,726 B · 2.441 MiB · MD5 18d654e4e560854540c75c8b3e2ed235		Beh_sup_train1_after_learning.npy113,352,887 B · 108.102 MiB · MD5 08c7066c45c4edfb871ff29b314ba64a
VR2_2021_04_11_12021-04-11 · block 1
sup_test1 · behavior key VR2_2021_04_11_1
Gender="Male"mname="VR2"datexp="2021_04_11"blk="1"sess#=1exptype="sup"rewType="Passive"stim_id=[0,1,2,3]is2p=1isDR=0artLick=1ROIdir=[]
sup_train2_before_learning · behavior key VR2_2021_04_11_1
Gender="Male"mname="VR2"datexp="2021_04_11"blk="1"sess#=1exptype="sup"rewType="Passive"stim_id=[0,1,2,3]is2p=1isDR=0artLick=1ROIdir=[]
VR2_2021_04_11_1_neural_data.npy
8,902,809,881 B · 8.291 GiB · MD5 781f17bc5f905297dfcc57024510cd73		VR2_2021_04_11_1_SVD_dec.npy
168,568,866 B · 160.760 MiB · MD5 d8d7a08e8a5d06e2307e18eb1dc88f6b		VR2_2021_04_11_trans.npz
3,046,050 B · 2.905 MiB · MD5 62c2430741056730b34918d9187b4e5f		Beh_sup_test1.npy180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c
Beh_sup_train2_before_learning.npy180,666,800 B · 172.297 MiB · MD5 f8f6c3d78fb17b386d9c9746e4cf821c
VR2_2021_04_28_12021-04-28 · block 1
sup_train2_after_learning · behavior key VR2_2021_04_28_1
Gender="Male"mname="VR2"datexp="2021_04_28"blk="1"exptype="sup"rewType="Passive"stim_id=[0,2,3]days=9is2p=1isDR=0artLick=1ROIdir=[]
VR2_2021_04_28_1_neural_data.npy
4,865,629,229 B · 4.531 GiB · MD5 52c39507f4470c5103dbc536a671c7fd		VR2_2021_04_28_1_SVD_dec.npy
132,700,062 B · 126.553 MiB · MD5 198dbd32e6e7a3cadf58e37ba7b30ad0		VR2_2021_04_28_trans.npz
2,557,610 B · 2.439 MiB · MD5 cf77d4b837cdab38917163a5666b097f		Beh_sup_train2_after_learning.npy141,228,534 B · 134.686 MiB · MD5 a1e4bcb573c7b1646daf8830648f4ef6
VR2_2021_04_29_12021-04-29 · block 1
sup_test2 · behavior key VR2_2021_04_29_1
Gender="Male"mname="VR2"datexp="2021_04_29"blk="1"sess#=1exptype="sup"rewType="Passive"stim_id=[0,2,3,4]is2p=1isDR=0artLick=1ROIdir=[]
VR2_2021_04_29_1_neural_data.npy
6,687,478,457 B · 6.228 GiB · MD5 8096285d55050570dc37965bb82f459a		VR2_2021_04_29_1_SVD_dec.npy
145,437,662 B · 138.700 MiB · MD5 4eeda94c7bb7a3de5fd671cd742e72d3		VR2_2021_04_29_trans.npz
2,613,210 B · 2.492 MiB · MD5 605e378ce517b34a574d856cffa17fb0		Beh_sup_test2.npy161,695,939 B · 154.205 MiB · MD5 67d5413083b3ffbb5f0b5ceddfd1c66b
VR2_2021_05_04_12021-05-04 · block 1
sup_test3 · behavior key VR2_2021_05_04_1_swap1
Gender="Male"mname="VR2"datexp="2021_05_04"blk="1"sess#=1stimtype="swap1"exptype="sup"rewType="Passive"stim_id=[0,2,5,3]is2p=1artLick=1ROIdir=[]
VR2_2021_05_04_1_neural_data.npy
4,125,446,273 B · 3.842 GiB · MD5 3ef852fb14b49426cb293cdc0a31aaa2		VR2_2021_05_04_1_SVD_dec.npy
118,698,462 B · 113.200 MiB · MD5 8afd1057e6d9ba9080cc5cdd6470cc95		VR2_2021_05_04_trans.npz
2,004,642 B · 1.912 MiB · MD5 b9b0936e4a1a3e899204a4cdcde375c6		Beh_sup_test3.npy169,187,762 B · 161.350 MiB · MD5 c9f3060140564dc56518c6f56a4683bb
VR2_2021_05_06_12021-05-06 · block 1
sup_test3 · behavior key VR2_2021_05_06_1_swap2
Gender="Male"mname="VR2"datexp="2021_05_06"blk="1"sess#=1stimtype="swap2"exptype="sup"rewType="Passive"stim_id=[0,2,6,3]is2p=1artLick=1ROIdir=[]
VR2_2021_05_06_1_neural_data.npy
5,301,160,673 B · 4.937 GiB · MD5 00ce52cfc42c322aea0165d0cea93324		VR2_2021_05_06_1_SVD_dec.npy
150,464,866 B · 143.494 MiB · MD5 2df8c62f80e522c63e6c9307df31f662		VR2_2021_05_06_trans.npz
2,765,142 B · 2.637 MiB · MD5 963a7e1de1667d0be57c7e91396cc7db		Beh_sup_test3.npy169,187,762 B · 161.350 MiB · MD5 c9f3060140564dc56518c6f56a4683bb
Audit provenance

Inventory JSON SHA-256: 4f498b97501d038a83ab6ba52bfec109e3b540019bbd37602e9382eb600c3d14. Experiment-index JSON SHA-256: ad8eaf217b3908976a3f701d6700d9ffd4479c529d8d2eab345919d694b57650. Source Imaging_Exp_info.npy MD5: 2259b9e5a6cea8987d871c7fbe90a8f9, SHA-256: 69756688e06fd60843b8aff988c2191f7719ec5a66d3660aee550a2ee2a26c56. Direct file URLs, exact bytes, and MD5 values above come from the deposited release manifest.

Back to contents ↑

07
Behavior and retinotopy coverage

What accompanies every full-neural acquisition—and what does not

Every one of the 89 full-neural acquisitions has a matching SVD file, matching retinotopy, at least one experiment membership, and a resolvable experiment-level behavior bundle in the release inventory. The modalities differ in grain and must be joined deliberately.

Modality semantics and scientifically relevant fields
Layer / grainReleased contentHow it enters analysis
Full neural
physical acquisition		Pickled dict with plane-wise spks arrays concatenated on the neuron axis. These are Suite2p-processed, deconvolved activity traces—not raw movies (Methods: processing of calcium imaging data ↗).Per-neuron paper estimator, distribution shape, thresholds, and full-versus-SVD benchmarks.
Reduced neural
physical acquisition		U is 400 × neuron and V is 400 × frame; reconstruct U.T @ V. It is lossy and the basis sees the full session.Fast exploration, population projections, and team-scale prototyping. Validate tails and spread against full traces before confirmation.
Retinotopy
mouse–date acquisition		iarea, xy_t, and position/transform fields map neurons to V1, medial, lateral, anterior, or excluded cortex.Area masks and density maps. It does not identify the same neuron across dates.
Imaging behavior
experiment bundle → session key → frame/trial		Trial IDs/times/walls/stimulus mapping/reward/lick/run and aligned fields including ft_trInd, ft_WallID, ft_Pos, ft_move, ft_CorrSpc, ft_RunSpeed, cue/reward/lick event frames, and position support.Role mapping, valid-frame mask, pairing/windowing, movement and occupancy QC, cue/reward/lick controls, and behavior–neural alignment.
Complete behavior-field glossary from the released processing notebook

The upstream data_process_script.ipynb in Drive documents a calcium-frame rate of 3.17 Hz and the schema below. Individual experiment bundles can contain protocol-specific missing or empty fields; the glossary is not permission to impute them.

Behavior dictionary keys, grouped without discarding any documented field
GroupFieldsMeaning / use
Trial identity and timingntrials, trInd, trInd_odd, trInd_even, Trial_start_time, Trial_end_timeTrial count, ordered indices, odd/even split, and corridor entry/exit times.
Stimulus identitystim_id, WallType, WallIsProbe, WallName, UniqWalls, TrialStim, StimTrial, StimFramestim_id defines roles 0–6. The upstream glossary explicitly says not to use generic WallType for these experiments and marks WallIsProbe as a catch-trial field to ignore here.
Subject movementSubjMove: SubjMTime, SubjMPos, SubjMPosCum, SubjMDistCum, SubjM_pitch, SubjM_roll, SubjM_yaw, SubjM_pitch_cumRaw movement timestamps, VR/cumulative positions, and ball-motion axes. Pitch-cumulative distance is not identical to VR cumulative position because the VR advances at fixed speed after threshold crossing.
VR position and epochsGray_space_time, VRpos, VRposCum, VRposTimeGrey-space entry and continuous VR position/time signals.
Cue and rewardSoundPos, SoundTime, SoundTimeDelay, RewTime, RewPos, isRewSound position/time, delayed sound time, actual reward time/position, and rewarded-trial indicator. Reward values are valid only where the protocol delivered reward.
LickingLickTrind, LickTime, LickPos, Lick_wallNameTrial, time, position, and stimulus identity for every lick.
Neural-frame alignmentft, ft_trInd, ft_trInd_odd, ft_trInd_even, ft_Pos, ft_PosCum, ft_move, ft_isMoving, ft_GraySpc, ft_CorrSpc, ft_WallID, ft_RunCum, ft_RunSpeed, RunFr, run_posPer-neural-frame time, trial, position, movement, corridor/grey mask, wall, and running values; run_pos is the trial × position speed derivative.
Event-frame indicesRewardFr, StartFr, GrayFr, EndFr, LickFr, SoundFr, SoundDelayFr, SoundDelPos, BefCueFr, AftCueFrNeural-frame indices/masks for reward, corridor entry, grey entry, exit, licks, cue/delayed cue, and before/after-cue epochs.
Protocol settingsCorridor_Length, Gray_Space_length, Texture_Length, Reward_Mode, Reward_Delay_msSession geometry and configured reward mode/delay. Use RewTime for actual delivery time.
Retinotopy-field glossary and area code
  • iarea: one integer label per neuron. The released loader maps 8→V1; {0,1,2,9}→medial HVA; {5,6}→lateral HVA; {3,4}→anterior HVA; −1 and 7 are excluded from those visual-cortical groups.
  • xy_t: transformed cortical coordinates aligned neuron-for-neuron with iarea. The paper code’s Get_density_map lines 394–416 ↗ uses x = −xy_t[:,1] and y = xy_t[:,0] before rasterization; Figure 1i ↗ is the plotted density-map output.
  • Some downstream caches/notebook products expose coordinate aliases xpos/ypos; the canonical paper loader consumes xy_t and iarea, so analysis should not require aliases that are not verified in a selected file.
  • areas.npz['out'] is a separate release-level set of area-outline polygons used for maps; it is not a per-acquisition neuron table.
  • Retinotopy assigns location/area inside an acquisition. It supplies no cross-date cell-registration map and cannot support neuron-identity persistence claims.
Requested missing-modality audit
Indexed IDs without full neural

0

Full IDs outside the index

0

Full IDs without SVD

0

Indexed dates without retinotopy

0

Metadata rows without full / retino

0 / 0

Experiment labels without behavior bundle

0

Therefore item 3 has no true imaging instance in the released index: there is no confirmed later behavioral/retinotopy session from an imaging mouse that lacks full neural data. Apparent extras are logical reuse of an acquisition, not modality-poor acquisitions. The static index proves each membership resolves to a published bundle; it does not claim that every internal pickle key was exhaustively opened.

Why some acquisitions appear more than once

Sixty-four acquisitions have one experiment label; 14 have two; 6 have three; 2 have four; and 3 have five. Those 44 extra logical memberships explain the difference between 89 acquisitions and 133 unique experiment–recording pairs. Nine pairs also contain two swap behavior instances:

Experiment–recording pairs with separate swap1 and swap2 behavior keys
ExperimentAcquisitionSession variants
naive_test3TX119_2023_12_13_1swap1 · swap2
naive_test3TX123_2023_12_18_1swap1 · swap2
naive_test3TX123_2023_12_20_1swap1 · swap2
naive_test3TX124_2023_12_24_1swap1 · swap2
naive_test3TX124_2023_12_26_1swap1 · swap2
unsup_test3DR10_2022_07_30_1swap1 · swap2
unsup_test3TX119_2024_01_09_1swap1 · swap2
unsup_test3TX123_2024_01_17_1swap1 · swap2
unsup_test3TX88_2022_07_22_1swap1 · swap2
Separate behavior-only cohort

The paper identifies 23 additional behavior-only mice implanted with headbars but no cranial windows (Methods: animals ↗) and reports their learning result in Figure 5 ↗. The three deposited behavior bundles total 1,698,037,349 bytes (1.581 GiB); the absence of neural/retinotopy joins is established by the Figshare v2 inventory, not inferred from Figure 5. The 23 imaging behavior bundles total 5,298,108,789 bytes (4.934 GiB); repeated atlas links are deduplicated by deposited file ID.

Protocol snapshots are not missing-file errors

  • TX108/TX109 have naive baselines before supervised training; TX119/TX123 have naive baselines before unrewarded exposure.
  • LZ13/LZ16/TX139 supply grating before/after acquisitions; TX140 and TX124 are naive-only.
  • TX61 enters the supervised series at Test 1 / Train 2-before; TX83 stops at unrewarded Test 1; TX104 and TX85 stop after Train 1.
  • Supervised Test 3 may use separate physical sessions for swap variants, whereas several naive/unrewarded Test 3 acquisitions hold two logical behavior instances in one neural acquisition.

Back to contents ↑

08
The environment

Workspace layout, drive.py, and graph.py

The dataset is large enough that a single, verified copy is shared rather than duplicated per user. Two small modules sit on top of that copy: drive.py, which provides safe access to the files, and graph.py, which turns ordinary Python functions into an editable, visible dataflow so that analyses can be run without editing code.

Google Drive workspace

The shared workspace folder holds the code and notebooks at its root; the dataset is a separate shared folder, added as a shortcut and kept read-only, whose layout mirrors the Figshare release. Members add the workspace as a shortcut in their own Drive; a notebook mounts Drive and imports the two modules from that path.

Zhong et al. 2025 — Workspace            (shared folder, added as a My Drive shortcut)
├── drive.py                            access layer over the release
├── graph.py                            analysis-graph runner and widget
├── zhong2025/                          catalog + relationship helpers (Python package)
├── notebooks/                          Colab notebooks (00, 03, 04, …)
├── README.md, pyproject.toml
│
└── Janelia dataset (Figshare v2)       separate shared folder, read-only
    ├── data/
    │   ├── beh/                        behaviour bundles + Imaging_Exp_info.npy
    │   ├── spk/                        full neural recordings
    │   ├── SVD_dec/                    400-PC reduced neural
    │   └── retinotopy/                 transforms + areas.npz
    └── metadata/                       catalog.csv · RELEASE.json · MD5SUMS · TRANSFER_STATUS.json

drive.py — access layer

The module exposes a short public workflow and keeps path resolution, checksum verification, pickle handling, caching and a size guard internal. A catalog drives lookups, so listing or filtering files never scans Drive. Two modes are supported: a connected mode that reads the mounted release and validates it against metadata/catalog.csv, RELEASE.json and TRANSFER_STATUS.json; and a metadata-only mode that uses the bundled catalog without fetching files.

Public data-access workflow
CallResult
data = setup()a Dataset (connected if the release is mounted, else metadata-only)
data.recording(id)a Recording, resolving its behaviour, SVD, full-neural and retinotopy files
data.recordings(experiment=, mouse=)recordings selected by scientific identity, not by path
data.find(category=, contains=, …)catalog rows matching a query, without opening any file
rec.load(layer) / data.load(name)verified, in-memory arrays or dictionaries (size and MD5 checked before load)
data.figshare() · data.picker()the Figshare article metadata · a visual file selector

Each catalog row (DataFile) carries a name, category, size, MD5 and the recording/retinotopy/experiment keys used to relate files; a Recording groups the files belonging to one recording. A default 1 GiB guard keeps the full-neural files an explicit choice, since every reduced-neural, behaviour and retinotopy file is well under that limit.

graph.py — analysis graph and widget

An analysis is expressed as ordinary functions. Decorating a function with @graph.node(outputs=…) makes it a node; its parameter names are input ports. A parameter that matches an upstream node’s output name is wired (a filled port); a parameter with a default value is an editable input (a hollow port). graph.Graph(name, *nodes) assembles the nodes, which execute one at a time in declaration order — the graph is acyclic by construction, so results are reproducible.

graph.widget(controls=…, show=…) renders the graph as a canvas: hollow ports become dropdowns and sliders, a Run flow button executes the nodes in order, and the wires trace data to the displayed output. The result of a run is available as panel.last_run, recording the settings, intermediate outputs and timings. The module is intentionally minimal — node, Graph, and run / run_many / diagram / widget — with no scheduler, persistence, parallelism or cross-run cache.

Verified Drive map

The current source set is in the shared workspace; the 421.175 GiB release is in the separate read-only dataset folder. The two paper PDFs used to substantiate this document are the Nature paper in Drive and the Science methods review in Drive. The Neuromatch planning document provides project context but is not primary evidence for scientific claims.

All current analysis notebooks in the shared Drive
NotebookOpenCorrect role and limit
00Data accessMount the shared release, inspect the catalog, choose files, and verify the selective cache. Access/QC only; no scientific estimator.
02Visual learning / RQ1 sandboxGraph 4 explores paired leaf–circle d′ distributions and SD/IQR/skew/excess-kurtosis/tails. It runs one of four example sessions at a time and is not the all-mouse confirmatory analysis.
03Dataset walkthroughCohort, retinotopy, alignment, and compact cross-validated population-d′ mechanics. Use for orientation and QC.
04Paper companionMaps protocol, paper figures, file semantics, and code provenance. It is a design reference, not an estimator.
05Reward / d′ dynamicsRQ2 workflow with preflight, held-out blockwise d′, slopes, saturation, cross-temporal and position analyses, exact mouse permutation, bootstrap, and leave-one-mouse-out checks.
06Older within-session demoOne rewarded and one unrewarded session with a cumulative in-sample estimator. Keep only as a sanity check; notebook 05 supersedes it.
ReferenceUpstream Neuromatch notebookProvenance/reference material, not the project’s confirmatory workflow.
Drive code capability map
ModuleWhat it provides—and does not provide
drive.pyCanonical 297-file catalog, recording/layer resolution, MD5 and size checks, atomic VM cache copy, disk check, and a 1 GiB default file guard. No HTTP/API fallback is implemented.
graph.pyStable sequential node/port execution and notebook widget. It is an orchestration/UI layer, not a statistical method, scheduler, persistent cache, or exporter.
zhong2025/learning.pyRQ2 toolbox: SVD contrasts, contiguous cross-validation, blockwise d′, position surfaces, cross-temporal matrices, slopes/saturation, exact permutation, mouse bootstrap, area transforms, and simulations.
zhong2025/data.pyPublic Figshare metadata and declared small download profiles. It deliberately does not provide an unrestricted full-neural network fallback.
zhong2025/position.pyBehavior–neural frame alignment and trial × position binning without interpolation across trials.
zhong2025/atlas.pyRelease inventory, experiment semantics, and complete recording-bundle resolution.
zhong2025/demo.pyCompact TX119 mechanics used for demonstrations, not group evidence.
Original pipelinePaper-provenance source, including utils.py and figure scripts. The processing notebook has large workstation-scale storage/runtime requirements and is not the team Colab path.
Graph stability contract

graph.py runs declared nodes sequentially and exposes a visible run-state card. Scientific correctness still belongs in pure tested functions. Every graph should show loading, success, empty-result, or the full error; disable duplicate runs; store settings, outputs, and timings in panel.last_run; and provide a non-widget callable path for tests and batch execution. Notebook 02 in Drive is a clean output-free copy.

09
Analysis methods for large-scale recordings

A methodological contract for this dataset

This section separates methods reported by Zhong et al. from safeguards proposed for new analyses. The large-scale-recording principles are sourced to Stringer & Pachitariu, Science 386, eadp7429 (2024); the reviewed copy is in the shared Drive. The paper-specific estimators remain linked to the exact Zhong Methods subsections.

Published review figures

The four complete figures and their published captions below are cropped from pages 3–6 of the Science review PDF in the shared Drive.

Science Figure 1, with panels A through C on single-neuron properties, population averaging, and population smoothing, followed by the published caption
Science Figure 1 — Single-neuron analyses at scale. Complete published figure and caption, cropped from PDF page 3 of the reviewed Drive PDF; artwork, labels, and plotted values are unchanged. Exact PDF page ↗
Science Figure 2, with panels A through D on geometry, topology, dynamics, and exploratory dimensionality reduction, followed by the published caption
Science Figure 2 — Population vectors and dimensionality reduction. Complete published figure and caption, cropped from PDF page 4 of the reviewed Drive PDF; artwork, labels, and plotted values are unchanged. Exact PDF page ↗
Science Figure 3, with panels A through C on encoding, decoding, and cross-validation, followed by the published caption
Science Figure 3 — Encoding, decoding, and cross-validation. Complete published figure and caption, cropped from PDF page 5 of the reviewed Drive PDF; artwork, labels, and plotted values are unchanged. Exact PDF page ↗
Science Figure 4 showing the review framework for single-neuron, population-structure, encoding, and decoding analyses, followed by the published caption
Science Figure 4 — Frameworks for analyzing large-scale recordings. Complete published figure and caption, cropped from PDF page 6 of the reviewed Drive PDF; artwork, labels, and plotted values are unchanged. Exact PDF page ↗

Published methods and proposed extensions

Published estimator, proposed extension, and validation source
StageAnalysisValidation criterion
1 · Reproduce the published estimatorCan the signed selectivity and regional fractions be reproduced?Use the paper’s frame selection, d′ definition and threshold exactly as reported (Neural selectivity ↗; Figure 1f,i–j ↗).
2 · Single-neuron distributionsHow do selectivity scale, asymmetry and tails vary by area, condition, stage and trial window?Label this as a new analysis; preserve the mouse/session hierarchy and use independent observations for learned selection or ordering (large-scale-recording review ↗).
3 · Population structureDoes a held-out population axis separate leaf and circle more strongly or earlier?Fit normalization and axes on training observations and score held-out temporal blocks; compare invariant summaries across mice (review ↗).
4 · Behavior and stateAre neural trajectories associated with licking, running, position, cue or reward timing?Keep sensory d′, late-versus-early cue-position d′ and chronological slope as distinct estimands (Sensory selectivity ↗; Reward-prediction neurons ↗).
Analysis safeguards from the methods review
  1. Write the biological estimand first. Separate sensory leaf–circle d′, late-versus-early cue-position reward-prediction d′, and chronological learning slope; the same symbol does not make them interchangeable.
  2. Cross-validate every learned choice. Neuron selection, sorting, coding axes, normalization learned from data, and decoders are fitted on training blocks and scored on unseen blocks.
  3. Block time. Slowly varying arousal, pose, running, and neural activity can create nonsense correlations under random-frame splits; use contiguous trial blocks (review).
  4. Keep mice as independent units. Millions of neuron-window values improve descriptive precision but do not turn 4 and 9 mice into a large cohort.
  5. Do not equate decoding with mechanism. With large populations many variables are decodable; perturbation or stronger design is needed for causality.
  6. Label exploration. The review supports visualization and dimensionality reduction for discovery, but confirmation needs locked metrics, independent scoring, and multiplicity control.

Back to contents ↑

10
The metric

One name, three different estimands

The paper quantifies per-neuron discriminability with d′ (definition and endpoint analysis: Nature Fig. 1f,i,jpaper ↗). Its means and standard deviations pool valid deconvolved-activity frames across trials for each corridor, using the original non-interpolated traces in the moving 0–4 m textured segment (Neural selectivity Methods).

Article definition

d′paper = (μ₁ − μ₂) / (σ₁/2 + σ₂/2)

Equivalently, d′ = 2(μ₁−μ₂)/(σ₁+σ₂). This is the paper-specific average-SD definition, not the pooled-variance effect-size formula.

Released implementation

2·(nanmean(x₁) − nanmean(x₂)) / (nanstd(x₁) + nanstd(x₂))

utils.py, lines 370–374 uses ddof=0. Positive values indicate stronger responses to the first argument.

Why one leaf trial + one circle trial needs a definition

If each trial is first collapsed to one scalar response, there is only one observation per stimulus and each within-stimulus standard deviation is undefined: this cannot be a classical d′. Retaining all frames makes the paper’s algebra numerically computable for one pair, but those frames are autocorrelated and have unequal position occupancy. That quantity may be displayed as a pairwise frame index, never presented as an independent inferential d′.

Trial-resolved specification

Make a deterministic sequence of nearest, unused leaf1–circle1 trial pairs. Around pair k, use a balanced rolling window containing a prespecified number K of pairs (default 8; sensitivity 5 and 10). For every neuron and trial, average valid activity within equal corridor-position bins and then give the bins equal weight. The window statistic is:

Windowed trial-level d′

d′n,w = 2( r̄n,leaf,w − r̄n,circle,w ) / ( sn,leaf,w + sn,circle,w )

Here the means and sample SDs (ddof=1) are across the K equal-position trial summaries for each role. A zero denominator, inadequate role count, excessive pair gap, or incomplete position support produces an explicit missing value and QC reason.

paper endpoint
d′paper: per-neuron, frame-level whole-session selectivity; reproduce this first with full traces, exactly as in Nature Fig. 1paper ↗
pair sandbox
pairwise frame index: one leaf + one circle trial, useful for a fast graph but descriptive only; show frame counts, position support and trial gap beside it
project primary
d′window: per-neuron balanced rolling-window selectivity based on trial-level, equal-position responses; this is a proposed extension, not a published statistic
valid frames
(ft_move>0) & ft_CorrSpc — the running, original-trace, 0–4 m texture mask specified in Methods: neural selectivity ↗
sign convention
first argument = role 2, leaf1; second = role 0, circle1. Positive values therefore mean leaf1-pole selectivity, regardless of the physical texture family
paper threshold
|d′paper| ≥ 0.3; the endpoint fraction is shown across thresholds 0.1–1.0 (Extended Data Fig. 1epaper ↗). For d′window, threshold tail mass is secondary and must be accompanied by a threshold sweep.
11
Proposed extension · not performed in the paper

Candidate question: how does the d′ distribution change?

Research question

Across ordered trial windows, how do the scale, asymmetry, and tail structure of the across-neuron leaf1-versus-circle1 d′ distribution change, and do those trajectories differ by Train 1 stage, supervised versus unrewarded exposure, and visual area?

The endpoint version asks how a distribution differs before versus after training; the trial-resolved version asks how it evolves inside each sampled session. Both are supported. A continuous neural day-by-day trajectory and same-cell change across dates are not supported by the released design (Nature Fig. 1bpaper ↗).

Testable hypotheses

Separate dimensions of distributional change
HypothesisPre-specified measuresWhat would support it
H1 · broader selectivityPrimary: log SD. Robust checks: IQR and MADSpread increases across progress or from before to after, especially in medial HVAs, where the paper found the largest selective-fraction increase (Fig. 1i,jpaper ↗).
H2 · directional asymmetryBias-corrected signed skewness; positive and negative tail fractionsSkewness changes only if one stimulus pole grows preferentially. Symmetric growth of leaf- and circle-selective tails can widen the distribution while skewness stays near zero.
H3 · selective-tail redistributionFisher excess kurtosis, |d′| quantiles, |d′| tail mass, and optional multimodality diagnosticsCentral mass moves into selective tails. No direction is preclaimed for kurtosis: rare extremes can raise it, whereas broad or bimodal separation can lower it.
Replace “flatter” or “stumpier” with the measured property

A larger σ means broader; it does not by itself imply lower standardized kurtosis, greater skewness, or heavier tails. SD/IQR/MAD measure spread, skewness measures directional asymmetry, excess kurtosis measures standardized tail/peak shape, and |d′|≥0.3 measures tail mass. Report them separately rather than using one visual adjective for all four.

The four requested strata

Train 1 design linked to the published timeline and endpoint comparisonpaper ↗
StratumRelease labelMiceRecordingsRole in the analysis
Supervised beforesup_train1_before_learning44Pre-learning task-cohort session; initial reward mode is passive
Supervised aftersup_train1_after_learning44Same mice 9–24 days later; reward mode is not uniform across all sessions
Unrewarded beforeunsup_train1_before_learning99Pre-exposure session with no reward or water restriction
Unrewarded afterunsup_train1_after_learning99Same mice 3–12 days later

Proposed primary area: medial HVAs, because Figure 1j reports the largest familiar-selectivity changes in that grouping (Results ↗; Figure 1i–j ↗). This is a design choice for the new distribution analysis, not a primary endpoint designated by the paper. V1, lateral HVA, and anterior HVA remain separate regional analyses. The anterior late-versus-early cue-position statistic is evaluated separately (Results: reward prediction ↗; Figure 4 ↗).

Relationship to the published result

The paper measured whole-session d′, selective-neuron fractions, and cortical density before versus after exposure (Fig. 1f–jpaper ↗; Extended Data Fig. 1paper ↗). The proposed distribution-shape and trial-window analyses extend those measurements; they were not reported by the authors. A result may therefore be paper-compatible without being a replication of a published panel.

Nature Fig. 1jpaper ↗ supplies the biological precedent and regional expectation. The complete pairing, graph, descriptive, and inferential protocol is given in §15.

12
Research question 2

Does reward alter within-session neural learning rate?

Research question 2

Does rewarded training change the early within-session rate of cross-validated leaf1-versus-circle1 neural discriminability relative to the unrewarded-exposure cohort, after accounting for trial support, movement, position, cue timing, and licking?

Proposed analysis—not a paper result. The paper reports familiar-stimulus selectivity increases after both rewarded training and unrewarded natural-texture exposure (Results: supervised and unsupervised plasticity ↗; Figure 1i–j ↗). It separately reports a late-versus-early cue signal in anterior HVAs of task mice (Results: reward prediction ↗; Figure 4e–g ↗). RQ2 asks a new chronological-rate question and does not restate either published endpoint.

Proposed RQ2 design
ElementPrimary specificationReason
Comparison subsetsup_train1_before_learning (4 passive-reward mice) versus unsup_train1_before_learning (9 no-reward mice)Avoids later reward-mode heterogeneity; this remains a cohort comparison, not isolated reward randomization.
Primary representationTemporally blocked, held-out population leaf–circle discriminability in fixed non-overlapping trial blocksAvoids same-trial fitting/scoring and overlapping-window pseudo-replication (methods review).
Primary areaPre-register medial HVA for sensory separation; V1/lateral/anterior are secondaryFigure 1j shows the largest reported familiar-selectivity fraction changes in the medial grouping (Figure 1j ↗); choosing it as the RQ2 primary is a proposed design decision.
Mouse estimandOne prespecified early slope per mouse; secondary time-to-threshold and saturation parametersKeeps inference at the independent unit and makes “faster” explicit.
Group testExact label permutation of mouse slopes; mouse/cluster bootstrap interval; leave-one-mouse-out stabilityAppropriate to n=4 versus n=9 and robust to one influential mouse.
Paper-reproduction checkReproduce the anterior d′late-versus-early reward-prediction result separatelyThe exact estimator and held-out selection procedure are in Methods: reward-prediction neurons ↗; its regional result is in Figure 4f–g ↗.

Hypotheses and interpretation

  • H2a · equal rates: rewarded and unrewarded mice have similar early slopes; this is compatible with exposure-driven medial plasticity.
  • H2b · reward-associated acceleration: rewarded mice have a larger early slope after the locked QC and nuisance sensitivities.
  • H2c · proposed regional dissociation: test whether the new sensory-rate effect concentrates in medial HVA; separately require reproduction of the paper’s anterior reward-prediction effect (Figure 4f–g ↗).
  • H2d · state explanation: an apparent group slope attenuates when speed, position support, cue/reward timing, and licking are examined jointly. This does not mean the state variables are mere nuisance if they are on the task pathway.
Causal-language limit

The imaging task protocol couples water restriction, sound cue, licking contingency, and reward delivery (Methods: water restriction ↗; Methods: reward delivery and lick detection ↗; Methods: behavioural training ↗). Therefore a cohort slope difference is reward-associated evidence, not an isolated reward manipulation. Figure 5 uses a separate behavior-only design (Figure 5a–b ↗) and cannot supply the missing neural causal contrast.

Back to contents ↑

13
Evidence boundaries

Published results and scope limits

The paper reports endpoint comparisons and a separate task-associated cue-position signal. It does not report the new chronological distribution and learning-rate analyses defined in Research question 1 and Research question 2. The table below keeps those evidential boundaries explicit.

Published evidence and scope limits
TopicWhat the paper reportsWhat is not established
Familiar-stimulus selectivityAfter Train 1, task and unrewarded natural-texture cohorts increased familiar-stimulus selectivity; the grating cohort did not show a comparable increase (Results: supervised and unsupervised plasticity ↗; Figure 1i–j ↗).The reported panels do not estimate a chronological within-session trajectory of the full signed d′ distribution.
Regional distributionThe medial grouping—PM, AM, MMA and lateral retrosplenial cortex—showed the largest reported familiar-selectivity changes (Results ↗; Figure 1i–j ↗).The endpoint comparison does not by itself identify whether a distribution changed in scale, asymmetry, tails, or mixture structure.
Task-associated signalNeurons meeting the late-versus-early cue-position criterion were concentrated primarily in anterior visual areas of task mice after training; the paper reports cue-, lick-, region-, running- and licking-related controls (Results: reward prediction ↗; Methods: reward-prediction neurons ↗; Figure 4d–l ↗).This statistic is not leaf1-versus-circle1 sensory d′, and the cohort design does not isolate reward from every task-associated variable.
Later behavioral learningIn 23 separate behavior-only mice, natural-texture pretraining was followed by faster discrimination learning than grating or no pretraining (Results: faster task learning ↗; Figure 5 ↗; Methods: animals ↗).Those mice had no cranial windows, so Figure 5 does not provide a neural learning-rate measurement.

Back to contents ↑

14
Published figure provenance

Paper panels, Methods, and code locations

This table maps published panels to the exact Results or Methods subsection and to the paper-tagged analysis code. It describes the published computation; proposed trial-window extensions remain in the research-question and analysis sections.

Published figure, method, and code map
Published panelsReported contentExact methodPaper code
Figure 1a–b ↗Virtual-reality task, cohort structure and training/test timeline.Visual stimuli ↗ · Behavioral training ↗Imaging_Exp_info.npy ↗ supplies the released session labels.
Figure 1f,i–j ↗Signed leaf1-versus-circle1 d′, selective-neuron density maps and regional fractions.Neural selectivity ↗dprime and density-map helpers ↗
Figure 2d–j ↗Preferred-position sequence comparisons, coding-direction projections and similarity indices.Neuron selection and sorting ↗ · Coding direction and similarity index ↗coding-direction and split-data helpers ↗
Figure 3a–h ↗Novel-stimulus selectivity, adaptation, leaf1-versus-leaf2 selectivity and coding-direction projections.Results: novelty and orthogonalization ↗ · Coding direction ↗coding-direction and selective-neuron helpers ↗
Figure 4d–l ↗Late-versus-early cue-position d′, regional distributions, cue/lick alignment and lick-versus-no-lick comparisons.Results: reward prediction ↗ · Reward-prediction neurons ↗reward-prediction helpers ↗
Figure 5a–h ↗Behavior-only pretraining cohorts, task structure, learning curves, first-lick positions and trial counts.Results: faster task learning ↗ · Behavioral training ↗ · Animals ↗The neural-imaging release does not supply recordings for these 23 behavior-only mice.

Back to contents ↑

15
Protocol · graph · inference

Complete analysis specification

1. Cohort and file manifest

Use all 26 Train 1 before/after recordings from the 13 matched mice in §05. Store one immutable manifest row per session with mouse, condition, stage, recording ID, date, reward mode, exact behaviour/full/SVD/retinotopy filenames, byte sizes and MD5 values. Develop against the 3.910 GiB SVD path, but run the final single-neuron distribution analysis on full deconvolved traces because those underlie the paper’s d′ result (Nature Fig. 1f–jpaper ↗).

2. Frame alignment and trial responses

  1. Align without guessing. Load one recording at a time; truncate neural and ft_* arrays to their shared frame count and fail visibly when identities or lengths disagree.
  2. Resolve roles per session. Read leaf1 as role 2 and circle1 as role 0 from stim_id; never infer the role from a literal texture filename.
  3. Apply the published support mask. Keep (ft_move>0) & ft_CorrSpc, matching the original-trace, running, 0–4 m estimator in Methods: neural selectivity ↗.
  4. Equalize corridor support for the extension. Bin the 0–4 m corridor into 20 equal bins, average frames inside each trial×bin, require a predeclared coverage fraction, then weight occupied bins equally. This stops slow or over-sampled positions from dominating a trial.
  5. Keep selection separate. Analyze every finite neuron in each pre-defined area. Do not select “responsive” neurons on the same window whose skewness or tails will be tested.

3. Trial pairing and balanced windows

Walk physical trial order and pair each trial with the nearest unused opposite role; break ties by the earlier trial, record both trial IDs, use the midpoint as pair time, and reject pairs beyond a fixed maximum gap (default 10 trials). Normalize midpoint rank to session progress 0–1. The interactive display may show the one-pair frame index, but the primary d′window uses complete rolling windows of eight pairs (five and ten as locked sensitivities). For confirmatory slopes, use non-overlapping eight-pair windows; overlapping windows are visualization only unless uncertainty is obtained with a block bootstrap.

4. Across-neuron distributions

For each mouse × session × area × window, save the neuron count and the following without pooling mice:

Distribution summary table
DimensionStatisticsImplementation
Locationmean and median signed d′Retain sign; also report median |d′|
SpreadSD, IQR, MADPrimary endpoint = log(SD); add a small predeclared epsilon only if needed
Asymmetrysigned skewnessscipy.stats.skew(..., bias=False, nan_policy="omit")
Tail / peak shapeFisher excess kurtosisscipy.stats.kurtosis(..., fisher=True, bias=False, nan_policy="omit")
Quantiles5th, 25th, 50th, 75th, 95th signed; 90th and 95th |d′|Show quantile differences, not only kernel densities
Selective tailsd′≥0.3, d′≤−0.3, and |d′|≥0.3 fractionsLink the threshold to Extended Data Fig. 1epaper ↗; repeat across 0.1–1.0
Quality supportrole trials, pair gap, frames/role, occupied bins, finite neurons, speed, dateNever silently drop a failed window

5. Interactive graph

Controls

Condition, stage, area, mouse, full/SVD representation, signed/absolute view, window size, maximum pair gap, position-bin count, minimum coverage, and tail threshold. Every control value is copied into the output manifest.

Panel A · distribution

ECDF plus violin or ridgeline by ordered window. ECDF is the stable default; density bandwidth is an explicit control. Never imply that pooled neurons are independent mice.

Panel B · shape trajectory

SD/IQR/MAD, skewness, excess kurtosis, |d′| quantiles and tail fractions over normalized progress. Draw every mouse lightly and group summaries strongly.

Panel C/D · inference & QC

Before/after mouse slopes and cohort contrasts beside trial gaps, support, speed and failures. A run-state card must show loading, success, empty-result, or the full error; no button press may fail silently.

Graph outputs must be deterministic for identical inputs: stable trial sorting and tie-breaking, explicit random seeds for bootstrap draws, no hidden mutable cache, and a fresh figure object per run. The displayed title includes recording ID, representation, window size, area, finite-neuron count and QC status so a stale result is visible.

6. Mouse-level inference

Primary area and metric: medial HVA and log SD. Estimate one non-overlapping-window progress slope per session, subtract before from after within each mouse, then compare those 13 mouse changes between the four supervised and nine unrewarded mice. Report the supervised-minus-unrewarded change, an exact cohort-label permutation p-value, a mouse-bootstrap 95% interval, and leave-one-mouse-out estimates. This two-stage analysis remains interpretable if a mixed model is singular.

Mixed-model sensitivity: on the mouse×session×non-overlapping-window table, fit log_sd ~ progress * C(stage) * C(condition) with a mouse random intercept and progress slope. The key term is progress:C(stage):C(condition). In statsmodels.formula.api: mixedlm(..., groups=data["mouse"], re_formula="~progress"). With only four supervised mice, report convergence, random-effect singularity, effect sizes and intervals; never treat a failed fit as a null result.

Secondary families: IQR/MAD, skewness, kurtosis, tail metrics, V1/lateral/anterior areas, and area interactions. Control false discovery rate within declared families, or present simultaneous mouse-bootstrap intervals. A binomial model is appropriate only for selective-neuron counts with the area neuron count as denominator; continuous skewness or kurtosis belongs in a continuous model or mouse-level permutation analysis.

7. Validation and sensitivity

  1. Reproduce the anchor. On full traces, recover whole-session d′paper, signed poles, |d′|≥0.3 fractions and the before/after medial-HVA direction from Nature Fig. 1paper ↗ before interpreting the new trajectory.
  2. Measure representation drift. Compare full and SVD single-neuron distributions on a declared session subset; report correlation and changes in SD, skewness, kurtosis and tails rather than assuming equivalence.
  3. Vary design choices one at a time. Repeat with K=5 and 10 pairs, alternative maximum gaps, 10/20/40 position bins, non-overlapping versus rolling windows, and the paper’s raw-frame estimator.
  4. Probe behaviour and acquisition. Plot speed, occupancy, frame count, cue/reward/lick timing and acquisition date beside neural metrics; use them as labelled sensitivities rather than automatically adjusting away condition effects.
  5. Preserve the hierarchy. Bootstrap mice, and retain both sessions and their windows with each sampled mouse. Do not bootstrap neurons as though they were independent animals.
16
Executable analysis runbook

How to use every notebook and code layer for both questions

This runbook separates what is executable now from what still needs implementation. Use the shared Drive workspace and one shortcut to the read-only release; do not make a 421 GiB copy per teammate.

Notebook order and evidential role
OpenUse
00 Data accessMount the shared release, inspect the catalog, choose files, and verify the selective cache. Access/QC only; no scientific estimator.
02 Visual learning / RQ1 sandboxGraph 4 explores paired leaf–circle d′ distributions and SD/IQR/skew/excess-kurtosis/tails. It runs one of four example sessions at a time and is not the all-mouse confirmatory analysis.
03 Dataset walkthroughCohort, retinotopy, alignment, and compact cross-validated population-d′ mechanics. Use for orientation and QC.
04 Paper companionMaps protocol, paper figures, file semantics, and code provenance. It is a design reference, not an estimator.
05 Reward / d′ dynamicsRQ2 workflow with preflight, held-out blockwise d′, slopes, saturation, cross-temporal and position analyses, exact mouse permutation, bootstrap, and leave-one-mouse-out checks.
06 Older within-session demoOne rewarded and one unrewarded session with a cumulative in-sample estimator. Keep only as a sanity check; notebook 05 supersedes it.
Reference Upstream Neuromatch notebookProvenance/reference material, not the project’s confirmatory workflow.

Execution order

  1. Access and provenance. Run notebook 00; record catalog/release checksums, selected file IDs, sizes, MD5 values, runtime package versions, and the immutable JSON analysis specification.
  2. Protocol and QC. Read notebook 04, then run notebook 03. Verify behavior–neural frame alignment, role mapping, valid-frame mask, trial counts, area counts, speed/position support, cue/reward/lick events, and exclusions by recording ID.
  3. Reproduce the paper anchor. On full traces for declared sessions, recover whole-session d′, signed poles, |d′|≥0.3 fractions, and the medial before/after direction from Nature Fig. 1paper ↗.
  4. RQ1 exploration. Use notebook 02 Graph 4 for a small number of sessions. State that its default 40 PCs / ≤2,000 neurons and frame-pair estimator are exploratory, not the final estimator.
  5. RQ1 confirmation—implementation still required. Add a deterministic maximum-lag trial pairer; trial × equal-position neuron summaries; fixed K-pair windows; all-neuron SD/MAD/IQR/skew/excess-kurtosis/quantile/tail metrics; a 4-stratum all-mouse runner; mouse-level inference; area multiplicity; and CSV/Parquet + JSON provenance export.
  6. RQ2. In notebook 05 run plan/preflight, mechanics, and simulations; explicitly switch from the default 2-per-group preview to all eligible 4+9 mice; run primary slope, exact permutation, clustered/mouse bootstrap, leave-one-mouse-out, cross-temporal, and position/behavior sensitivities.
  7. Representation validation. Compare 400-component/all-neuron SVD results with full traces on representative sessions. The paper’s single-neuron estimator uses full deconvolved traces; SVD is a computational approximation.
  8. Persist everything. Save the locked spec, checksums, per-session QC, per-mouse tidy metrics, model/permutation/bootstrap output, figures, exclusions, and software environment. panel.last_run alone is not a durable scientific artifact.
Exact analysis products for RQ1
  • rq1_spec.json: cohort, area primary, role order, K, pair rule/max lag, position bins, valid mask, minimum support, metrics, multiplicity families, and sensitivities.
  • rq1_window_metrics.parquet: mouse, acquisition, experiment, stage, condition, area, window, progress, counts/gaps/support, SD/MAD/IQR/skew/kurtosis/quantiles/tails, and QC reason.
  • rq1_mouse_effects.csv: one before→after and within-session effect per mouse/area/metric, with the locked primary medial log-SD effect.
  • Figures: ridgeline/ECDF/quantile curves, metric trajectories with mouse lines, signed tail plots, and QC panels. Density bandwidth never substitutes for distribution statistics.
Exact analysis products for RQ2
  • rq2_spec.json: Train 1-before cohorts, primary area, contiguous folds, block size/horizon, slope definition, exclusions, exact permutation, bootstrap, LOMO, and nuisance sensitivities.
  • rq2_mouse_slopes.csv: one primary slope per mouse plus saturation/time-to-threshold and QC summaries.
  • rq2_cross_temporal.zarr or compact array artifact: training-block × test-block generalization, with position-specific summaries.
  • Separate files for the anterior late-versus-early reward-prediction reproduction check; never merge that statistic with sensory leaf–circle d′.

Team-scale data access and fallback

The RQ1 26-session SVD/behavior/retinotopy plan is about 3.910 GiB; its full-neural sources are 126.080 GiB. The RQ2 before-only SVD plan is about 1.94 GiB; full sources are about 62 GiB. Each Colab runtime currently copies selected Drive files to ephemeral local cache, so concurrent users multiply shared-file reads. Google states that Colab resources and limits fluctuate, Drive mounts can hit per-user/per-file operation and bandwidth quotas, and popular shared files are a typical trigger.

An API fallback should not make every teammate download the same multi-GiB full files again. The stable design is: mounted Drive → checksum-addressed object-store/CDN derivative → exact public Figshare file → MD5 verification → local VM cache, with bounded retries and jittered exponential backoff for 403/429 responses (official Drive API quota guidance). Run full-neural preprocessing centrally once and publish compact trial × position or metric artifacts; keep team Colab fallback to reduced/compact derivatives.

Current implementation limits

Notebook 02 is not yet an all-mouse RQ1 analysis. Notebook 05 defaults to a 2-per-group preview, uses a session-wide SVD basis, has no full-neural estimator, treats behavior adjustment as diagnostic rather than causal modeling, and does not persist a complete result package automatically. The Drive workspace is also not a full source mirror; generator scripts and tests can differ from local copies.

Back to contents ↑

17
Released paper recipes

Recipe reference

The computations behind the paper-compatible recipes above, with direct links to the immutable paper release. These functions reproduce the publication’s frame-level analyses; the proposed balanced trial-window estimator is specified separately in §15.

Core primitives — d′, frame validity, roles, areas

d′(x1, x2)

utils.py:370–374 u = nanmean(x, axis=1); sig = nanstd(x, axis=1, ddof=0); d′ = 2·(u1 − u2) / (sig1 + sig2)

x1, x2 are neurons×frames; the caller groups the frames. The article and code both divide by the average of the two standard deviations, not a pooled root-mean-square variance. Positive d′ indicates tuning to the first argument.

Per-neuron d′ from the SVD without full reconstruction

derived from dprime + the 400-PC basis meanS = UT·V̄S ; meanSquareS = diag(UT·(VSVST/|S|)·U) ; d′ = 2(meanA−meanB)/(√varA+√varB)

The moments reproduce d′ for the 400-component SVD reconstruction without materializing the neurons×frames matrix. They do not make the lossy reconstruction identical to the full deconvolved traces used by the publication recipe.

Frame-validity mask

utils.py:431–433 fr_valid = (ft_move[:nfr] > 0) & ft_CorrSpc[:nfr]

Running and inside the 0–4 m texture. Always truncate ft_* to nfr = spk.shape[1].

get_cat_id(WallName, isRew) — physical wall → role

utils.py:137–147

The rewarded wall maps to role 2 (leaf1); its second exemplar to 3; the first unrewarded wall to 0 (circle1); its second exemplar to 1. This fixes leaf1 = role 2 and circle1 = role 0 for all rewarded cohorts and physical pairs. In unrewarded sessions isRew is uniformly false; read the roles from stim_id instead.

neu_area_ID(iarea) — area masks

utils.py:312–324 V1: iarea==8 · mHV: {0,1,2,9} · lHV: {5,6} · aHV: {3,4} · exclude {−1,7}

The medial selector is retinotopy[i]['neu_ar_idx']['mHV']. Identifiers −1 and 7 lie outside visual cortex and are dropped from density maps.

Selectivity — per-neuron d′, fractions, density maps

Get_dprime_selective_neuron(db, Beh, stim_ID=[2,0])

utils.py:418–441

Load the full {rec}_neural_data.npy traces; resolve stim1=UniqWalls[stim_id==2][0] and stim2=UniqWalls[stim_id==0][0]. dp = dprime(spk[:, (ft_WallID==stim1)&fr_valid], spk[:, (ft_WallID==stim2)&fr_valid]) — all valid frames, no odd/even split. Returns {'dprime':[per-session], 'retinotopy':[…], 'mname':[…]}.

Get_selective_neuron_fraction_with_dprime(dprime, retinotopy, dp_thrs=[0.1..1.0])

utils.py:443–457

Per area, per mouse, per threshold: both=|dp|≥thr, stim1=dp≥thr, stim2=dp≤−thr, each divided by the area neuron count. Output value has shape 4 areas × mice × n_thr × 3; index thr=0.3 → thr_idx=2; slice 0='both', 1=leaf1 pole, 2=circle1 pole.

Get_density_map(dprime, retinotopy, dp_thr=0.3, typ='both', n_down=10) + get_dist_img

utils.py:376–416

Mask by typ (both / stim1 / stim2). xpos=−xy_t[:,1], ypos=xy_t[:,0]. Rasterize into 5000² (+800 offset); define the footprint from the all-neuron background blurred with σ=1 and truncate=30; blur the selected-neuron image with σ=30; divide by neuron count; take the NaN-aware mouse mean; then downsample every tenth pixel. Preserve the helper’s left-right flip and the plotting orientation; render magma_r, vmax=5e-8, with areas.npz['out'] outlines.

Population — interpolation, coding direction, reward prediction

get_interpPos_spk / spk_pos_interp — 60-bin position tensor

utils.py:105–131

From spk[:, VRmove], ft_PosCum[VRmove], Corridor_Length → neurons × ntrials × 60. Bins ~0–40 are the textured metres, ~40–60 the grey space. Cached to {rec}_interpolate_spk.npy. Used by the coding-direction and reward-prediction recipes, not the primary d′.

Get_coding_direction(stim_ref=[2,0], prc=5, n_bef=10) + load_coding_direction

utils.py:503–597 · 349–364

d′ selection on ft_trInd%2==0 frames; corr_neu = texture>grey; normalize spk_norm = 2(spk−u0)/(std1+std2), u0 = grey bins 42:52; prepend 10 previous-trial grey bins; per area stim1_idx=dp≥pct95, stim2_idx=dp≤pct5. resp_diff = proj_2_stim1 − proj_2_stim2; proj_tr = resp_diff[:, 10:50].mean(1). The article’s similarity index is SI=(dx−dy)/(dx+dy); the code’s anchor-span denominator is equivalent when the target projection lies between the two anchors.

Get_dprime_rewPred_neuron(dp_thr=0.3) — anterior value d′

utils.py:459–501

sel_ind = stimulus-d′ ≥ 0; u_spk = interp_spk[:,:,5:40].mean(2); SoundPos = mod(SoundDelPos,60); dp1 = dprime(late-cue vs early-cue leaf1 trials); idx1 = (dp1≥0.3) & sel_ind; per-area fraction. This is a value/timing index in anterior HVAs, distinct from the medial stimulus d′.

get_lick_response_in_zone — behaviour metric

utils.py:280–295

lickResp = lickCount(def_range=[20,40])['inRange'] (lick in 2–4 m); rew_id = get_cat_id; istim in [2,0]; first 200 trials; perf.reshape(n_mice,5,2). Discrimination = −np.diff = reward − non-reward.

Key parameters

Methods

Paper constants, each with its exact location: |d′|≥0.3 and the running 0–4 m selectivity mask (neural-selectivity Methods ↗); 0.75 s deconvolution decay (calcium-processing Methods ↗); 6 cm s−1 run threshold and 4 m texture + 2 m grey corridor (visual-stimulus Methods ↗); 0.5–3.5 m imaging cue and 2–3 m behavior-only reward-zone start (behavioural-training Methods ↗); 20,547–89,577 traces per recording (plasticity Results ↗); paired/independent two-sided Student’s t-tests and no multiplicity adjustment (statistics and reproducibility ↗). The 400-component SVD is a Figshare-v2 release derivative, not a paper Methods constant (release inventory ↗). The proposed distribution analysis instead uses the locked mouse-level hierarchy in §15.

Read the paper code before extending it

The paper-tagged repository is compact enough to audit directly. The snippets below are the parts most likely to change the scientific interpretation of a plot. They are shown with their observation unit, split logic, and project-safe extension rather than as unexplained syntax.

1 · The published d′ is frame basedutils.py 370–374
def dprime(x1, x2):              # neurons × frames
    u1, u2 = np.nanmean(x1, 1), np.nanmean(x2, 1)
    s1, s2 = np.nanstd(x1, 1), np.nanstd(x2, 1)  # ddof=0
    return 2 * (u1 - u2) / (s1 + s2)

What it means. Each neuron gets one signed contrast between all valid leaf-role frames and all valid circle-role frames. The denominator is the arithmetic mean of the two population SDs, not pooled RMS variance and not Cohen's d. Positive means the first argument. Reproduce this exact endpoint before interpreting a trial-resolved extension (Nature Fig. 1f–jpaper ↗).

2 · The caller fixes roles and valid framesutils.py 418–441
stim1 = uniqW[stim_id == 2][0]       # analysis role 2
stim2 = uniqW[stim_id == 0][0]       # analysis role 0
valid = (beh['ft_move'][:nfr] > 0) & beh['ft_CorrSpc'][:nfr]
dp = dprime(spk[:, (ft_WallID == stim1) & valid],
            spk[:, (ft_WallID == stim2) & valid])

Why it matters. “Leaf” and “circle” are role names derived from stim_id; physical textures differ between some cohorts. The published comparison also excludes stationary and grey-space frames. A plot that infers roles from filenames, forgets the common frame truncation, or silently includes grey space is a different estimand.

3 · Selection, ordering, and display use different datautils.py 599–703
# first trial parity: estimate d′ and select neurons
train_frames = (ft_trInd % 2 == 0) & valid
dp = dprime(spk[:, stim1 & train_frames], spk[:, stim2 & train_frames])

# other parity: split again into ordering and displayed responses
test_spk = selected_position_tensor[:, reference_trials & (trial % 2 == 1)]
order_data, display_data = test_spk[:, ::2], test_spk[:, 1::2]
sort_id = np.argsort(np.argmax(order_data.mean(1)[:, :40], axis=1))

What the paper protected against. The first physical-trial parity selects neurons. Within the other parity, one subset determines peak-position order and the complementary subset is displayed. The comments call zero-based parity “odd/even”, so use the literal modulo test, not the comment. This role separation underlies the sorted sequences in Nature Fig. 2paper ↗ and is documented in Methods: neural selectivity ↗.

4 · The density map is not a local selective fractionutils.py 376–416
selected = np.abs(dp) >= 0.3
x, y = -xy_t[visual, 1], xy_t[visual, 0]
image = gaussian_filter(raster(selected), sigma=30)
image = image / visual.sum()        # total visual-cortex neurons
group_map = np.nanmean(mouse_images, axis=0)

Interpretation. The plot is a smoothed selected-neuron spatial density divided by the total visual-cortex neuron count for each mouse, followed by a mouse mean. It is not selected / all neurons at each pixel. Bandwidth changes appearance but cannot establish that the d′ distribution broadened; use ECDFs, quantiles, and locked moments for that new claim (Figure 1i–jpaper ↗).

5 · Coding direction is a population contrastutils.py 503–597
# normalize each neuron relative to grey space and the two stimulus spreads
spk_norm = 2 * (spk_position - grey_mean) / (stim1_sd + stim2_sd)
leaf_axis   = spk_norm[leaf_selective].mean(0)
circle_axis = spk_norm[circle_selective].mean(0)
projection  = leaf_axis - circle_axis

Proposed extension. This averages selected populations and asks where each trial lies on a leaf–circle axis. It is complementary to a distribution of per-neuron d′ values: population separation can improve even if only a subset changes. For new analyses, learn normalization, selection and the axis on training blocks and score held-out trials (analysis safeguards ↑).

6 · Reward-prediction panels use held-out neuron selectionutils.py 814–882
tr_shuf = np.random.permutation(ntrials)
for fold in range(10):
    dp_stim = dprime(train_circle_cue, train_leaf_cue)
    dp_value = dprime(train_late_cue_leaf, train_early_cue_leaf)
    selected = anterior & (dp_stim > 0.3) & top_five_percent(dp_value)
    held_out_response[test_trials] = activity[selected][:, test_trials].mean(0)

Important boundary. This is a late-versus-early cue/value statistic in anterior HVA, not the sensory leaf–circle d′ trajectory. The function calls random.seed(2025) but shuffles with NumPy's RNG, so the published helper is not deterministically seeded by that line. New RQ2 code should use an explicit np.random.default_rng(seed) or, preferably, deterministic contiguous folds. Interpret beside Nature Fig. 4paper ↗ and ED Fig. 8paper ↗.

How the project code turns those recipes into stable analyses

Paper recipe → tested project extension → most useful plot
NeedImplementationPlot and decision
Trial-level responses without crossing trial boundarieszhong2025/position.py bins each physical trial by corridor position and leaves unsupported bins missing; it does not extrapolate across trials.Trial × position support heatmap. Stop or mark a window missing when one role lacks the declared position coverage.
Trial-window d′learning.py 51–78 uses sample SDs (ddof=1) across trial summaries; 444–521 builds blockwise sensory contrasts.Small-multiple trajectories by area and stratum, with thin mouse lines and a mouse-level summary. Rolling windows are descriptive; non-overlapping blocks support inference.
Fast SVD momentslearning.py 154–234 obtains per-neuron means/variances from the reduced basis without reconstructing the full matrix.SVD-versus-full Bland–Altman/scatter on representative sessions. Do not claim exact single-neuron tails until agreement is measured.
Held-out population scorelearning.py 298–412 assigns contiguous folds before filtering, learns standardization and direction on other folds, and never pools raw scores across independently fit folds.Fold-wise d′ trajectory under the analysis safeguards; invalid folds remain visible with the support reason.
Stability across time and positionlearning.py 524–647 and 650–721 compute position surfaces, slopes, saturation, and cross-temporal generalization.Train-block × test-block heatmap and trial-progress × position surface. Broad off-diagonal generalization supports a stable axis; diagonal-only structure suggests transient state.
Independent-unit inferencelearning.py 724–860 collapses to unique mice, enumerates all 715 assignments for the 4-versus-9 comparison, and bootstraps mice.Mouse slope dot plot, exact permutation distribution, bootstrap interval, and leave-one-mouse-out forest plot.

Plot playbook for the two questions

Use each plot for one clearly stated job
PlotAnswersGuardrail
ECDF or signed quantile curvesDid the whole d′ distribution translate, widen, or change asymmetrically?Show signed values; same x-axis across strata; add mouse-resampled uncertainty rather than treating neurons as independent.
Ridgeline / violin / KDEWhere is the mass, and is a tail or second mode visually plausible?Keep one common bandwidth and axis; treat it as visualization, not the estimator of SD or kurtosis.
Metric trajectory small multiplesHow do log SD, IQR/MAD, skew, excess kurtosis, and signed tail fractions evolve over trial progress?One facet per metric, area, and declared stratum; thin mouse paths; never hide a dimension behind a toggle in the confirmatory figure.
Signed tail plotAre leaf- and circle-selective tails recruited symmetrically?Plot P(d′≥t) and P(d′≤−t) together over a threshold sweep, anchored to ED Fig. 1e–fpaper ↗.
Cross-temporal matrixDoes an axis learned early still separate stimuli later, or only within the same block?Fit every row on its training block, test every column without refitting, and annotate class support (analysis safeguards).
QC companion panelCould speed, occupancy, cue position, reward, licking, or unequal role counts explain the trajectory?Show the QC next to the result, not in an appendix. Use ED Fig. 2paper ↗ and ED Fig. 8paper ↗ as paper precedents.
Minimal statistics code for exploration, followed by the inferential boundary
from scipy import stats

summary = {
    "sd": np.std(dp, ddof=1),
    "iqr": stats.iqr(dp, nan_policy="omit"),
    "skew": stats.skew(dp, bias=False, nan_policy="omit"),
    "excess_kurtosis": stats.kurtosis(dp, fisher=True, bias=False,
                                      nan_policy="omit"),
    "leaf_tail": np.mean(dp >= threshold),
    "circle_tail": np.mean(dp <= -threshold),
}

# Secondary hierarchical model after one row per mouse/session/window is built.
model = smf.mixedlm(
    "log_sd ~ progress * stage * condition",
    metric_rows,
    groups=metric_rows["mouse"],
    vc_formula={"session": "0 + C(session)"},
)

The SciPy dictionary is descriptive and should be computed for every declared stratum. The mixed model is secondary because the cohort is small and trajectories are correlated. RQ1's primary effect should be a prespecified mouse-level before→after change; RQ2's 4-versus-9 primary comparison should use the exact mouse-label permutation, with bootstrap and leave-one-mouse-out uncertainty. Neuron-window rows are never the independent sample size.

18
Constraints

Constraints and interpretation rules

  • Published anchor versus new analysis. Whole-session d′, |d′|≥0.3 fractions, and their before/after regional pattern were published (Nature Fig. 1paper ↗). Trial-window distribution trajectories, skewness and kurtosis are proposed extensions and must never be described as paper results.
  • Sparse dates, not daily imaging. The release samples protocol milestones. Train 1 pairs are 3–24 days apart, so “day-by-day neural learning” is unsupported even though trials are ordered inside every recording (Fig. 1bpaper ↗).
  • No same-cell longitudinal claim. The release supplies cortical retinotopy for every recording but no cross-date neuron identity map. Pair before/after sessions by mouse; do not subtract neuron i on one date from neuron i on another.
  • One-pair terminology. One scalar leaf response and one scalar circle response cannot estimate two within-role SDs. A frame-based one-pair index is descriptive and autocorrelated; d′window requires multiple balanced trials per role.
  • Window dependence. Adjacent rolling windows reuse trials and all windows reuse the session’s neurons. Use them for smooth visualization, not as independent replicates; confirm with non-overlapping windows or block resampling.
  • Mouse-level inference. Four supervised and nine unrewarded mice are the independent cohort units. Twenty-six sessions, thousands of trials, and millions of neuron-window values do not change those group sample sizes.
  • Representation matters. The 400-component SVD is a lossy exploratory representation. Full deconvolved traces are required for exact paper-style single-neuron distributions, and SVD-to-full agreement must be measured for spread and tails.
  • Cohort wording. “Supervised” is a task/reward-present protocol, not one uniform action-contingent intervention; after-learning reward modes vary. Reward condition is also bundled with water restriction, motivation, licking, and history, so an interaction is a cohort difference rather than automatic proof of reward causality.
  • Fixed sign and physical roles. Positive d′ always means role 2 over role 0 after session-specific role resolution. Because mice can see different physical texture families, do not infer sign from a filename or switch to |d′| before inspecting pole asymmetry.
  • Distinct shape claims. Greater SD does not entail greater skewness or kurtosis. Choose medial HVA and log SD before inference; treat other metrics, areas, thresholds, bandwidths, and multimodality checks as multiplicity-controlled secondary families.
  • Support and confounds. Role counts, pair gaps, frame counts, corridor occupancy, speed, missing bins, cue/reward/lick timing, acquisition date and finite-neuron counts must be visible beside trajectories. Empty or invalid windows remain in the audit table with a reason.
  • Control limits. The non-significant lick comparison in Nature Fig. 4lpaper ↗ checks one explanation in another test context; it neither establishes causation nor excludes reward and state effects here.
  • Behaviour-only data are separate. Nature Fig. 5paper ↗ and Extended Data Fig. 9paper ↗ provide daily behaviour from 23 additional mice but no neural recordings; they cannot fill the imaging gaps.
  • Null-result language. Similar or uncertain distribution trajectories mean that the specified change was not detected at this sample size and resolution. They do not prove that learning or reinforcement leaves selectivity unchanged.

Clean report order: paper-anchor reproduction; all-mouse distribution plots; primary medial-HVA log-SD effect and uncertainty; within-mouse before/after slopes; cohort contrast; shape and area families with multiplicity control; QC and representation validation; then limitations beside the result.