Choice-specific sequences in parietal cortex during a virtual-navigation decision task

Choice-specific sequences in parietal cortex during a virtual-navigation decision task Harvey Coen Tank

Tag layer of the DHSS Meta-Tagging Project. The article text is reproduced from the staged copy named in sourceDesc and is not altered; the tags are held separately in standOff so that overlapping and nested spans are preserved.

Choice-specific sequences in parietal cortex during a virtual-navigation decision task 2012 Nature 10.1038/nature10918 Harvey2012-choice-specific-sequences-in-parietal-cortex Harvey2012-choice-specific-sequences-in-parietal-cortex.txt, the project's staged copy 25ae31b810cd076d7f29490eb036f02c52822982 not-redistributable neuroscience

the paper's own words

the tag layer; a sentence written by the tagger about the paper, always anchored to a quotation but not itself a quotation

0 spans located by searching the stored text for the quote's exact words, allowing any whitespace between them the temporal and anatomical structure of population activity in mouse posterior parietal cortex during a memory-guided choice Quotation withheld because this record is not cleared for redistribution. whether PPC dynamics are better described as classes of cells with sustained activity or as high-dimensional sequences Quotation withheld because this record is not cleared for redistribution. shows choice-specific sequences spanning the whole trial, implemented by cells anatomically intermixed below 100 microns, and argues this replaces the stable-state account Quotation withheld because this record is not cleared for redistribution. existence Quotation withheld because this record is not cleared for redistribution. choice-specific neuronal sequences in posterior parietal cortex Quotation withheld because this record is not cleared for redistribution. electrophysiology electrophysiology Quotation withheld because this record is not cleared for redistribution. mechanisms A decision is reached when the evolving activity sequence intersects one of the choice-specific trajectories in population state space Quotation withheld because this record is not cleared for redistribution. mechanisms Growing state-space separation between the two choice trajectories over the trial progressively raises the barrier to reversing a decision Quotation withheld because this record is not cleared for redistribution. mechanisms The parietal cortex switches between sustained-activity and sequence-based dynamical regimes according to the spatiotemporal demands of the behavioral task Quotation withheld because this record is not cleared for redistribution. information moves from one neuronal population to another across the trial, each cell active for only a fraction of it Quotation withheld because this record is not cleared for redistribution. a proposed framework: the sequence starts choice-independent, then moves toward one choice-specific trajectory, and the growing state-space distance between trajectories acts as a barrier to changing the decision Quotation withheld because this record is not cleared for redistribution. choice can be decoded from about 65 neurons within 250 microns, above chance in cue, delay and turn periods Quotation withheld because this record is not cleared for redistribution. functionally distinct subnetworks are anatomically interlaced, implying highly specific synaptic connectivity rather than spatial clustering Quotation withheld because this record is not cleared for redistribution. the mouse PPC is located first by retrograde and anterograde labelling, since the region had not been characterised in this species Quotation withheld because this record is not cleared for redistribution. necessity established by muscimol inactivation, with a visually-guided version of the task as the control for visual and motor deficits Quotation withheld because this record is not cleared for redistribution. trials aligned to cue offset, turn onset and trial end, because running speed and turn position varied across trials and mice Quotation withheld because this record is not cleared for redistribution. The task confounds the decision with ongoing navigation, so the decision process itself was not isolated. what this limits: every decision-related claim in the paper: OUR BEHAVIORAL PARADIGM DID NOT ISOLATE THE DECISION-MAKING PROCESS Quotation withheld because this record is not cleared for redistribution. Whether the sequences serve decision-related processes remains to be tested. what this limits: the functional role of the sequences: SEQUENCES MAY THEREFORE PLAY A ROLE IN SOME ASPECT OF DECISION-RELATED PROCESSES, BUT FURTHER EXPERIMENTS WILL BE NECESSARY Quotation withheld because this record is not cleared for redistribution. The weak distance-dependence of pairwise correlation may be an optical artefact rather than real spatial organisation. what this limits: the spatial-organisation result: the weak distance dependence of pairwise correlation may be an optical artefact rather than real spatial structure Quotation withheld because this record is not cleared for redistribution. The result may not generalise across PPC regions, layers or species. what this limits: the generality of the sequence result: DIFFERENT REGIONS OR LAYERS OF THE PPC MAY HAVE DIFFERING ACTIVITY DYNAMICS, OR THE DYNAMICS OF RODENT AND PRIMATE PPC CIRCUITS MAY DIFFER Quotation withheld because this record is not cleared for redistribution. the paradigm does not isolate decision-making, so the role of sequences in the decision itself is untested what this limits: any role for the sequences in the decision itself: OUR BEHAVIORAL PARADIGM DID NOT ISOLATE THE DECISION-MAKING PROCESS Quotation withheld because this record is not cleared for redistribution. the weak distance-correlation relationship may be overlapping fluorescence rather than real spatial organisation, and the paper declines to choose duplicate of: limitations[2] · what this limits: the same scope as limitations[2] - the weak distance-correlation relationship may be overlapping fluorescence rather than real spatial organisation Quotation withheld because this record is not cleared for redistribution. four rival explanations are offered for why this result differs from the primate literature and none is selected what this limits: the disagreement with the primate literature, which is left unresolved among four alternatives: SEQUENCES ... MAY BE PRESENT IN THE PRIMATE PPC DURING TRADITIONAL TASKS BUT HAVE YET TO BE IDENTIFIED Quotation withheld because this record is not cleared for redistribution. spatial and temporal progression through the task were highly correlated, so what the active cell encodes cannot be separated what this limits: what an active cell in the sequence encodes: NAVIGATION, MEMORY, AND CHOICE INFORMATION MAY BE COMBINED IN THE SEQUENCES, and spatial and temporal progression through the task are highly correlated Quotation withheld because this record is not cleared for redistribution. the mouse PPC is a stipulated region, defined by projection pattern and position rather than by any functional criterion what this limits: every use of the label 'mouse PPC' and every cross-species comparison built on it: BECAUSE THE LOCATION OF THE MOUSE PPC HAS NOT BEEN CHARACTERIZED, WE FIRST PERFORMED RETROGRADE AND ANTEROGRADE LABELING EXPERIMENTS TO LOCATE IT ANATOMICALLY ... CONSISTENT WITH THE RAT AND PRIMATE PPC Quotation withheld because this record is not cleared for redistribution. the state-space trajectory view of population activity, taken explicitly from work on behavioural choice in the leech nervous system Quotation withheld because this record is not cleared for redistribution. feedforward architectures and liquid state machines, offered as possible implementations Quotation withheld because this record is not cleared for redistribution. two versions of the virtual T-maze differing ONLY in the cue period and the reward location eliminates: everything the two mazes share - the visual environment, the running, the task structure - leaving the cue and the reward as the only thing a difference in neural activity can be about · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. sequence here means each neuron active for only a fraction of a task period with different cells active at different times - explicitly opposed to a class of cells sharing one time course Quotation withheld because this record is not cleared for redistribution. high-dimensional vs low-dimensional is used as the criterion separating the two accounts: low-dimensional means each cell's activity is an amplitude-scaled copy of one prototypical time series Quotation withheld because this record is not cleared for redistribution. delay cell, cue cell and response cell are treated as possibly artifactual categories - the paper finds the three groups but shows sequences inside each one Quotation withheld because this record is not cleared for redistribution. mouse PPC is a stipulated object: the region was located by its afferents, efferents and position relative to visual and somatosensory cortex, then declared to be PPC Quotation withheld because this record is not cleared for redistribution. neuronal circuit trajectory - the population state as a point in an n-dimensional space with one dimension per neuron Quotation withheld because this record is not cleared for redistribution. claim_type: existence a second reading reaching the same value. Recorded: existence — note: a second reading reaching the same value from a different sentence. Not applied - a person decides. Quotation withheld because this record is not cleared for redistribution. method: two-photon imaging of GCaMP3 in layer 2/3 PPC of head-restrained mice navigating a virtual T-maze, with muscimol inactivation and open-loop and linear-track controls a second reading DIFFERING from the recorded value. Recorded: electrophysiology — note: a second reading, verbatim-verified, disagreeing with the value on the record. Not applied - a person decides. Quotation withheld because this record is not cleared for redistribution. phenomenon: sequential population dynamics during working memory and decision a second reading DIFFERING from the recorded value. Recorded: choice-specific neuronal sequences in posterior parietal cortex — note: a second reading, verbatim-verified, disagreeing with the value on the record. Not applied - a person decides. Quotation withheld because this record is not cleared for redistribution. extends → previous work on spatial intermixing of response properties in sensory cortex, motor cortex and hippocampus (refs 12, 31–32) Quotation withheld because this record is not cleared for redistribution. contradicts → models predicting spatial clustering of similar response patterns and connectivity from axonal-dendritic overlap (refs 46–47) differs in: spatial clustering and connectivity rule: the opposing models EMPHASIZED THE SPATIAL CLUSTERING OF SIMILAR RESPONSE PATTERNS, AS IN FUNCTIONAL COLUMNS, AND PROPOSE CONNECTIVITY DEFINED BY AXONAL-DENDRITIC OVERLAP WITHOUT FINE-SCALE SPECIFICITY Quotation withheld because this record is not cleared for redistribution. builds-on → studies identifying cortical sequences of activity states during working memory tasks (refs 27–30) Quotation withheld because this record is not cleared for redistribution. extends → the trajectory account of behavioural choice first developed for the leech nervous system (ref. 37) Quotation withheld because this record is not cleared for redistribution. contradicts → the account of PPC as classes of cells with sustained activity spanning whole task periods differs in: homogeneous classes against sequences: CLASSES OF CELLS WITH HOMOGENEOUS ACTIVITY PATTERNS WERE NOT PRESENT. RATHER, CHOICE-SPECIFIC SEQUENCES OF NEURONS WERE ACTIVATED IN ALL BEHAVIORAL PERIODS Quotation withheld because this record is not cleared for redistribution. contradicts → models predicting spatial clustering of similar response patterns, as in functional columns differs in: the same parameter as item 1 - spatial clustering, AS IN FUNCTIONAL COLUMNS, against fine-scale specificity Quotation withheld because this record is not cleared for redistribution. extends → reports of cortical activity sequences during working memory in prefrontal cortex and hippocampus Quotation withheld because this record is not cleared for redistribution. distinguishes-from → hippocampal place cells, since PPC cells did not encode location on a linear track differs in: THE AXIS IS A SPECIFIC TASK THAT SEPARATES THEM. 'It seems unlikely that PPC neurons only provided location information in A CONTEXT-DEPENDENT MANNER, LIKE HIPPOCAMPAL PLACE CELLS, BECAUSE UNLIKE PLACE CELLS PPC NEURONS DID NOT ENCODE SPATIAL LOCATION DURING A LINEAR TRACK TASK.' The distinction is drawn by a task the alternative account predicts they would pass - so the axis names the experiment that discriminates, not a property of either population. Quotation withheld because this record is not cleared for redistribution. qualifies → drift-diffusion and recurrent winner-take-all models of decision-making differs in: what the dynamics settle into: DECISIONS ARE REACHED WHEN THE ACTIVITY APPROACHES A CHOICE-SPECIFIC FIXED POINT WITH WORKING MEMORY MAINTAINED AS STABLE ACTIVITY AT THAT POINT, and THE TRAJECTORIES ARE DEFINED BY RELATIVELY HOMOGENEOUS changes - against a sequence Quotation withheld because this record is not cleared for redistribution. reconciles → the primate literature reporting persistent activity, by proposing the task demands select the dynamics differs in: the task the animal was doing: THE PPC ADOPTS DIFFERENT DYNAMICS DEPENDING ON THE DEMANDS OF THE BEHAVIORAL TASK, which is offered as the explanation for HETEROGENEITY AND SEQUENCES IN OUR EXPERIMENTS VERSUS STEREOTYPY AND LOW-DIMENSIONAL DYNAMICS Quotation withheld because this record is not cleared for redistribution. low-dimensional —operationalised-as→ recurrent attractor network models low-dimensional recurrent attractor network models strength of the relation: 0.9 · note: the model is what makes the dimensionality claim testable - each cell an amplitude-scaled copy of one time series Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but low-dimensional dynamics 'CAN BE REPRODUCED in' recurrent attractor models; operationalised-as is loose. Quotation withheld because this record is not cleared for redistribution. choice-specific cells —part-of→ task-modulated cells choice-specific cells task-modulated cells strength of the relation: 1 Quotation withheld because this record is not cleared for redistribution. PPC —distinct-from→ place cells PPC place cells differs in: THE AXIS IS A NEGATIVE TEST USED TO RULE OUT AN ALTERNATIVE. 'It seems unlikely that PPC neurons only provided location information in A CONTEXT-DEPENDENT MANNER, LIKE HIPPOCAMPAL PLACE CELLS, BECAUSE UNLIKE PLACE CELLS PPC NEURONS DID NOT ENCODE SPATIAL LOCATION DURING A LINEAR TRACK TASK.' The comparison is not offered as a finding about PPC - it is offered as the reason a competing interpretation fails, so the axis is what the alternative would have predicted. · strength of the relation: 0.9 · note: argued from a negative result on a linear track, not from a positive dissociation Quotation withheld because this record is not cleared for redistribution. t-test t-test Quotation withheld because this record is not cleared for redistribution. principal component analysis principal component analysis Quotation withheld because this record is not cleared for redistribution. correlation correlation Quotation withheld because this record is not cleared for redistribution. rank correlation rank correlation Quotation withheld because this record is not cleared for redistribution. Kolmogorov-Smirnov test Kolmogorov-Smirnov test Quotation withheld because this record is not cleared for redistribution. multiple regression multiple regression Quotation withheld because this record is not cleared for redistribution. factor analysis / dimensionality reduction factor analysis / dimensionality reduction Quotation withheld because this record is not cleared for redistribution. classifier (decoding accuracy) classifier (decoding accuracy) Quotation withheld because this record is not cleared for redistribution. shuffle / permutation control shuffle / permutation control Quotation withheld because this record is not cleared for redistribution. Spearman's correlation of pair difference against distance Spearman's correlation of pair difference against distance Quotation withheld because this record is not cleared for redistribution. shuffled dataset control shuffled dataset control Quotation withheld because this record is not cleared for redistribution. multiple regression on running parameters multiple regression on running parameters Quotation withheld because this record is not cleared for redistribution. principal component analysis principal component analysis Quotation withheld because this record is not cleared for redistribution. Individual parietal neurons were each active only briefly, and their brief activations were staggered in time so that the population formed a sequence spanning the whole trial. Quotation withheld because this record is not cleared for redistribution. Different sequences of neurons ran on left- versus right-choice trials, making the sequences choice-specific. Quotation withheld because this record is not cleared for redistribution. Silencing the posterior parietal cortex with muscimol dropped memory-guided task accuracy to near chance without changing how many trials the mice ran. Quotation withheld because this record is not cleared for redistribution. The same inactivation did not significantly affect a visually-guided version of the task. Quotation withheld because this record is not cleared for redistribution. Cells preferring left versus right choices were anatomically intermingled rather than spatially segregated, since pair selectivity difference did not vary with inter-cell distance. Quotation withheld because this record is not cleared for redistribution. A small, anatomically local population of neurons was enough to decode the animal's upcoming choice above chance throughout the cue, delay and turn periods. Quotation withheld because this record is not cleared for redistribution. Differences in the animals' running trajectories between left and right trials could not account for the choice-specific activity. Quotation withheld because this record is not cleared for redistribution. The sequential ordering was not an artifact of sorting, because shuffled data produced no comparable sequences. Quotation withheld because this record is not cleared for redistribution. activity forms a sequence spanning the entire trial, with cells staggered in time Quotation withheld because this record is not cleared for redistribution. different sequences run on left and right choice trials Quotation withheld because this record is not cleared for redistribution. the sequences survive shuffling controls, so they are not an artifact of sorting cells by peak time Quotation withheld because this record is not cleared for redistribution. cells fall into cue, delay and turn groups, but each group contains a sequence rather than a class Quotation withheld because this record is not cleared for redistribution. no classes of cells with homogeneous activity patterns were present Quotation withheld because this record is not cleared for redistribution. left- and right-choice-preferring cells were anatomically intermixed Quotation withheld because this record is not cleared for redistribution. muscimol in PPC drops memory-guided performance to chance without affecting trial rate Quotation withheld because this record is not cleared for redistribution. the same inactivation leaves the visually-guided version of the task intact Quotation withheld because this record is not cleared for redistribution. running trajectory parameters could not explain the choice-specific activity Quotation withheld because this record is not cleared for redistribution. passive viewing and a linear track produced far less location-specific activity than the real task Quotation withheld because this record is not cleared for redistribution. individual trajectories switched between the mean correct left and right choice trajectories, most often during the delay Quotation withheld because this record is not cleared for redistribution. choice-specific cells choice-specific cells Quotation withheld because this record is not cleared for redistribution. neuronal circuit trajectory neuronal circuit trajectory Quotation withheld because this record is not cleared for redistribution. activity state of the circuit activity state of the circuit Quotation withheld because this record is not cleared for redistribution. t COM t COM Quotation withheld because this record is not cleared for redistribution. trajectory selectivity index trajectory selectivity index Quotation withheld because this record is not cleared for redistribution. neuronal circuit trajectory neuronal circuit trajectory Quotation withheld because this record is not cleared for redistribution. low-dimensional dynamics low-dimensional dynamics Quotation withheld because this record is not cleared for redistribution. task-modulated cells task-modulated cells Quotation withheld because this record is not cleared for redistribution. PPC inactivation PPC inactivation Quotation withheld because this record is not cleared for redistribution. neuronal activity neuronal activity Quotation withheld because this record is not cleared for redistribution. anatomical distance anatomical distance Quotation withheld because this record is not cleared for redistribution. trial-type selectivity trial-type selectivity Quotation withheld because this record is not cleared for redistribution. trajectory selectivity index trajectory selectivity index Quotation withheld because this record is not cleared for redistribution. 83 ± 9% — mean behavioral accuracy of mice on the memory-guided T-maze task Quotation withheld because this record is not cleared for redistribution. 54 ± 5% — fraction of trials correct after bilateral muscimol inactivation of PPC (near the 50% chance level) Quotation withheld because this record is not cleared for redistribution. ~71% — proportion of task-modulated cells that were choice-specific Quotation withheld because this record is not cleared for redistribution. n = 404 cells from 6 mice — total number of choice-specific, task-modulated cells pooled across animals Quotation withheld because this record is not cleared for redistribution. < 100 micrometers — the microcircuit length scale over which functionally distinct cells were anatomically intermixed Quotation withheld because this record is not cleared for redistribution. 63% of error trials, 20% of correct trials — share of trials on which the population trajectory switched selectivity during the delay or late cue period Quotation withheld because this record is not cleared for redistribution. Quotation withheld — fraction of neurons with significant location-specific activity increases in each of the three conditions Quotation withheld because this record is not cleared for redistribution. ~65 neurons separated by < 250 µm — size and anatomical spread of the neuronal population sufficient to decode choice above chance Quotation withheld because this record is not cleared for redistribution. 83 plus or minus 9% correct, p < 0.0001 vs 50% — behavioural accuracy Quotation withheld because this record is not cleared for redistribution. about 65 cells per field of view, range 37 to 94 — simultaneous imaging yield Quotation withheld because this record is not cleared for redistribution. about 73% of highly active cells were task-modulated — proportion carrying task signal Quotation withheld because this record is not cleared for redistribution. about 71% of task-modulated cells were choice-specific — proportion distinguishing left from right Quotation withheld because this record is not cleared for redistribution. 11 plus or minus 8% of trial time points with a transient — how briefly each cell is active Quotation withheld because this record is not cleared for redistribution. simulated runs 1.3%, linear track 5.8%, T-maze 32.3% — fraction of cells with location-specific activity, the control comparison Quotation withheld because this record is not cleared for redistribution. cue 5.1 plus or minus 2.6 s, delay 4.2 plus or minus 1.2 s, turn 3.0 plus or minus 1.8 s — task period durations Quotation withheld because this record is not cleared for redistribution. EVIDENCE CORRECTED (RE-GROUNDED - THE POINTER WAS WRONG, NOT THE CLAIM) · no injections 87 plus or minus 7%, saline 85 plus or minus 5%, muscimol 54 plus or minus 5% — the inactivation effect EVIDENCE CORRECTED: RE-GROUNDED - THE POINTER WAS WRONG, NOT THE CLAIM. The old evidence was verbatim, anchored and correctly cut, and it contained NONE of this claim's numbers (54, 85, 87). The paper states the claim in one sentence elsewhere and the tagger did not use it. the project, 2026-08-23: 'WE HAVE SPENT THE DAY VERIFYING THAT QUOTES ARE REAL AND NONE OF IT ASKS WHETHER THEY ARE RELEVANT.' Confirmed by reading the claim, the old span and the candidate together; the candidate carries every number AND the claim's own content words. Quotation withheld because this record is not cleared for redistribution.

The full text of this article is not redistributable, so it is not included here. The evidence sentences are withheld with it; the tag layer and its grounding status remain. The article itself is at https://doi.org/10.1038/nature10918.