Distinct relationships of parietal and prefrontal cortices to evidence accumulation

Distinct relationships of parietal and prefrontal cortices to evidence accumulation Hanks Kopec Brunton Duan Erlich Brody

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.

Distinct relationships of parietal and prefrontal cortices to evidence accumulation 2015 Nature 10.1038/nature14066 Hanks2015-distinct-relationships-of-parietal-and-prefr Hanks2015-distinct-relationships-of-parietal-and-prefr.txt, the project's staged copy 50f862799dc0192850f9d98b445db8d09136ada8 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 whether parietal and frontal cortex relate to accumulating evidence in the same way, tested by asking what each region's firing rate says about the value of the accumulator at each moment Quotation withheld because this record is not cleared for redistribution. whether frontal premotor activity participates in the graded accumulation of evidence or only in converting it into a discrete choice Quotation withheld because this record is not cleared for redistribution. parietal cortex encodes the accumulator as a graded value while frontal cortex encodes the sign of it, and silencing frontal cortex changes behaviour only at the end of the stimulus Quotation withheld because this record is not cleared for redistribution. mechanism Quotation withheld because this record is not cleared for redistribution. evidence accumulation in perceptual decision-making Quotation withheld because this record is not cleared for redistribution. electrophysiology electrophysiology Quotation withheld because this record is not cleared for redistribution. mechanisms the FOF does not integrate evidence; it reads out the sign of an accumulator computed upstream and converts it into a discrete motor plan at the moment the go signal could arrive, which is why perturbing it matters only at the end of the stimulus Quotation withheld because this record is not cleared for redistribution. mechanisms unilateral FOF silencing biases behavior by flipping already-categorized contralateral choices into ipsilateral ones, rather than by corrupting the accumulation itself Quotation withheld because this record is not cleared for redistribution. parietal firing answers what the value of the accumulated evidence is, through a graded map that is stable across time Quotation withheld because this record is not cleared for redistribution. frontal firing answers which choice the evidence so far favours, clustering into a positive and a negative group Quotation withheld because this record is not cleared for redistribution. the shorter click latency in frontal cortex suggests sensory information reaches it through circuits that bypass parietal cortex Quotation withheld because this record is not cleared for redistribution. the inactivation effect is modelled as converting a fraction of already-categorised contralateral choices into ipsilateral ones, which is a step after accumulation rather than during it duplicate of: mechanisms[1] Quotation withheld because this record is not cleared for redistribution. There is no primate frontal eye field dataset against which the click-triggered-average result can be compared. what this limits: the click-triggered-average result: NO EQUIVALENT EXPERIMENT FROM PRIMATE FEF EXISTS AS YET FOR COMPARISON Quotation withheld because this record is not cleared for redistribution. Data collection and analysis were not blinded to experimental group. what this limits: every group comparison in the paper: INVESTIGATORS WERE NOT BLINDED TO EXPERIMENTAL GROUPS DURING DATA COLLECTION OR ANALYSIS Quotation withheld because this record is not cleared for redistribution. The 500 ms and 250 ms optogenetic groups were run sequentially, so group assignment was not randomised. what this limits: the comparison between the two optogenetic durations: THE ASSIGNMENT OF GROUP WAS NOT RANDOMIZED - the groups were run sequentially Quotation withheld because this record is not cleared for redistribution. Two rats were dropped from the fine-timing optogenetic analyses on the basis of their full-trial effects. what this limits: the fine-timing optogenetic results: TWO RATS WERE NOT INCLUDED ... BECAUSE THEY DID NOT EXHIBIT SIGNIFICANT FULL-TRIAL EFFECTS - the exclusion criterion is the effect being measured Quotation withheld because this record is not cleared for redistribution. investigators were not blinded to experimental groups during data collection or analysis what this limits: the same scope as limitations[1] - unblinded collection and analysis Quotation withheld because this record is not cleared for redistribution. two rats were dropped from the finer-timing experiments because they had not shown a significant full-trial effect, and the two timing groups were run sequentially so group assignment was not randomized what this limits: the same scope as limitations[3] - selection on the outcome Quotation withheld because this record is not cleared for redistribution. the accumulator value used to build the tuning curves is estimated partly from the choice the animal went on to make what this limits: every accumulator tuning curve: A BETTER ESTIMATE OF THE ACCUMULATED EVIDENCE CAN BE OBTAINED BY ALSO TAKING INTO ACCOUNT THE CHOICE MADE AT THE END OF THE TRIAL - the regressor is informed by the outcome it is used to explain Quotation withheld because this record is not cleared for redistribution. the encoding difference does not survive the end of the decision, since the parietal map becomes categorical during motor preparation what this limits: the scope of the PPC/FOF encoding difference in time: THE DIFFERENCE IN ENCODING BETWEEN PPC AND FOF WAS NOT MAINTAINED AFTER THE END OF THE DECISION PROCESS Quotation withheld because this record is not cleared for redistribution. no comparable click-triggered measurement exists in the primate frontal eye fields for comparison what this limits: the cross-species comparison: no comparable click-triggered measurement exists in the primate frontal eye fields, so the analogy is not tested against one Quotation withheld because this record is not cleared for redistribution. preferred sides are mixed rather than uniformly contralateral, at about 59 percent in both areas what this limits: any contralateral-preference claim: THE PREFERRED SIDE WAS CONTRALATERAL FOR 59.1% OF THE PPC NEURONS AND 59.4% OF THE FOF NEURONS - mixed selectivity, not a uniform map Quotation withheld because this record is not cleared for redistribution. a fitted trial-by-trial accumulator, used as the independent variable that neural responses are plotted against Quotation withheld because this record is not cleared for redistribution. the click-triggered average, isolating one stimulus quantum's effect by subtracting the expected response for that Poisson rate Quotation withheld because this record is not cleared for redistribution. halorhodopsin silencing on a random quarter of trials, used to convert an encoding claim into a timing prediction Quotation withheld because this record is not cleared for redistribution. single-parameter variants of one behavioural model compared by bootstrapped likelihood ratio Quotation withheld because this record is not cleared for redistribution. MODEL-FREE analyses run alongside the model-based ones eliminates: the accumulation model itself as the source of the conclusion - if only the fitted model showed gradual accumulation, the finding could be a property of the model rather than of the rats · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. THE STANDARD ANALYSIS SAYS THE TWO AREAS ARE THE SAME AND A DIFFERENT ANALYSIS ON THE SAME DATA SAYS THEY ARE NOT: trial-averaged ramping looks alike in both, and the tuning curve assay separates a graded code from a categorical one Quotation withheld because this record is not cleared for redistribution. and the same firing rate is read as the answer to two different questions - what is the value of the accumulated evidence, against which choice should I make if the signal came now Quotation withheld because this record is not cleared for redistribution. THE RIVAL ACCOUNTS ARE SORTED BY WHEN THEY ACT: two of the four modelled biases must operate throughout accumulation and two only at its end, which is exactly what the timed silencing can separate Quotation withheld because this record is not cleared for redistribution. so the model comparison and the causal experiment are testing the same distinction in two ways rather than checking each other Quotation withheld because this record is not cleared for redistribution. THE STIMULUS IS DESIGNED SO THAT THE ANALYSIS IS VALID: because clicks come from a Poisson process, one click's timing says nothing about another's, which is what licenses the click-triggered average at all Quotation withheld because this record is not cleared for redistribution. THE LATENT VARIABLE IS INFERRED PARTLY FROM THE OUTCOME: the better estimate of accumulated evidence uses the choice the animal went on to make, and the frontal claim is about encoding that same choice Quotation withheld because this record is not cleared for redistribution. ENCODING IS NOT NECESSITY: parietal cortex carries a veridical graded code for the accumulator, and the same laboratory reports that its activity is not needed for the behaviour Quotation withheld because this record is not cleared for redistribution. the sample is declared unselected and then selected: every isolatable neuron was recorded regardless of response properties, and the analysis keeps the side-selective ones for a stated reason Quotation withheld because this record is not cleared for redistribution. and both disclosures are made - the recording area was assigned randomly, the timing groups were not Quotation withheld because this record is not cleared for redistribution. claim_type: mechanism a second reading reaching the same value. Recorded: mechanism — 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: tuning curves relating firing rate to a trial-by-trial estimate of the accumulator from a fitted behavioural model, compared between areas, and unilateral optogenetic silencing in four separate time windows 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: gradual accumulation of sensory evidence toward a binary choice, and its neural correlates in more than one cortical area a second reading DIFFERING from the recorded value. Recorded: evidence accumulation in perceptual decision-making — 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. contradicts → the prevailing view that prefrontal cortex is part of the neural circuit for evidence accumulation (refs 3, 5, 7, 8) Quotation withheld because this record is not cleared for redistribution. builds-on → the previously developed "Poisson Clicks" accumulation task (ref. 10) Quotation withheld because this record is not cleared for redistribution. replicates → previous findings of pre-movement side selectivity in rat PPC (refs 14, 15) and FOF (ref. 13) Quotation withheld because this record is not cleared for redistribution. contradicts → the prevailing view that prefrontal cortex is part of the circuit that accumulates evidence differs in: THE AXIS IS WHEN IN THE TRIAL THE PERTURBATION MATTERS. 'Unilateral FOF perturbations Quotation withheld, when the provisional choice must be converted into a motor act.' The prevailing view is not denied a role for prefrontal cortex - it is denied the ACCUMULATION role, on the evidence of WHEN the effect appears. Quotation withheld because this record is not cleared for redistribution. supports → primate work reporting ramping correlates of accumulation in parietal and frontal eye field neurons Quotation withheld because this record is not cleared for redistribution. extends → the graded parietal encoding seen in monkeys, which was measured in a task that did not require temporal accumulation differs in: THE AXIS IS WHETHER THE TASK REQUIRED TEMPORAL ACCUMULATION AT ALL. 'Quotation withheld.' The earlier finding is granted in full and its task is named as lacking the thing this paper is about - so a 'similar encoding' across species is a similarity of signal, not of computation, and only the task says which. Quotation withheld because this record is not cleared for redistribution. qualifies → the idea of a single distributed decision process shared across parietal and frontal areas Quotation withheld because this record is not cleared for redistribution. builds-on → the authors' own trial-by-trial behavioural model of the clicks task, which supplies the accumulator estimate Quotation withheld because this record is not cleared for redistribution. tuning curve —operationalised-as→ firing rate tuning curve firing rate strength of the relation: 1 · note: THE NEW ASSAY: a fitted behavioural model supplies a moment-by-moment estimate of the accumulator, and firing rate is plotted against it at each timepoint Quotation withheld because this record is not cleared for redistribution. graded —distinct-from→ categorical graded categorical strength of the relation: 1 · note: THE STANDARD ANALYSIS SAYS THE AREAS ARE ALIKE AND THE NEW ONE SAYS THEY ARE NOT - same neurons, same trials, opposite conclusions Quotation withheld because this record is not cleared for redistribution. graded —operationalised-as→ firing rate graded firing rate strength of the relation: 1 · note: the parietal rate answers what the value of the mentally accumulated evidence is, through a map that is stable across time Audit 2026-09-28: DOUBTFUL - read against its span and kept, but graded encoding was read from firing rates; operationalised-as is loose. Quotation withheld because this record is not cleared for redistribution. categorical —operationalised-as→ firing rate categorical firing rate strength of the relation: 1 · note: and the frontal rate answers a different question about the same quantity - which choice to make if the go signal came now Audit 2026-09-28: DOUBTFUL - read against its span and kept, but categorical encoding was read from firing rates; operationalised-as is loose. Quotation withheld because this record is not cleared for redistribution. steepness —operationalised-as→ tuning curve steepness tuning curve strength of the relation: 1 · note: and it holds per neuron as a distribution shift, with the two populations overlapping Audit 2026-09-28: DOUBTFUL - read against its span and kept, but steepness was computed per tuning curve; operationalised-as is loose. Quotation withheld because this record is not cleared for redistribution. click-triggered average —operationalised-as→ firing rate click-triggered average firing rate strength of the relation: 1 · note: one click has a sustained effect in both areas, decaying only in frontal cortex, and no primate frontal comparison exists Quotation withheld because this record is not cleared for redistribution. latency —operationalised-as→ click-triggered average latency click-triggered average strength of the relation: 1 · note: and the 200 against 100 ms difference is read as a bypassing route rather than as a processing delay Audit 2026-09-28: DOUBTFUL - read against its span and kept, but latency was measured on the click-triggered average; operationalised-as is loose. Quotation withheld because this record is not cleared for redistribution. side-selective —operationalised-as→ pre-movement period side-selective pre-movement period strength of the relation: 1 · note: 23 percent of parietal and 32 percent of frontal neurons qualify, kept because they are the ones most likely to matter for the decision Quotation withheld because this record is not cleared for redistribution. side-selective —operationalised-as→ firing rate side-selective firing rate strength of the relation: 1 · note: and in both areas those neurons ramp with a slope proportional to stimulus strength, which is the correlate the classical analysis finds Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but side-selective neurons showed ramping FIRING RATES; the side-selectivity itself is defined by a firing-rate difference, so operationalised-as is loose. Quotation withheld because this record is not cleared for redistribution. inactivation —causes→ bias inactivation bias strength of the relation: 1 · note: full-trial silencing gives a 10.3 percent ipsilateral bias Quotation withheld because this record is not cleared for redistribution. post-categorization —operationalised-as→ bias post-categorization bias strength of the relation: 1 · note: defined as flipping a fraction of already-categorised contralateral choices, with that fraction the only free parameter Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span defines the post-categorization bias; operationalised-as 'bias' is circular. Quotation withheld because this record is not cleared for redistribution. input gain —distinct-from→ input noise input gain input noise strength of the relation: 1 Quotation withheld because this record is not cleared for redistribution. graded —distinct-from→ choice graded choice strength of the relation: 1 · note: ENCODING IS NOT NECESSITY: parietal cortex carries a veridical graded code and is reported elsewhere not to be needed for the behaviour Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span says PPC is not necessary for choice; distinct-from is loose. Quotation withheld because this record is not cleared for redistribution. accumulator shift —contrasted-with→ input gain accumulator shift input gain differs in: THE AXIS IS WHEN IN THE ACCUMULATION THE MECHANISM ACTS. 'Quotation withheld.' Four candidate mechanisms fall into two groups by timing, and the grouping is what an experiment could test. Replaces the retracted within-group edge. · strength of the relation: 0.8 Quotation withheld because this record is not cleared for redistribution. ROC analysis ROC analysis Quotation withheld because this record is not cleared for redistribution. bootstrap bootstrap Quotation withheld because this record is not cleared for redistribution. Mann-Whitney U test Mann-Whitney U test Quotation withheld because this record is not cleared for redistribution. nonlinear curve fitting nonlinear curve fitting Quotation withheld because this record is not cleared for redistribution. log-likelihood model comparison log-likelihood model comparison Quotation withheld because this record is not cleared for redistribution. likelihood ratio likelihood ratio Quotation withheld because this record is not cleared for redistribution. Fano factor Fano factor Quotation withheld because this record is not cleared for redistribution. confidence interval confidence interval Quotation withheld because this record is not cleared for redistribution. standard error of the mean standard error of the mean Quotation withheld because this record is not cleared for redistribution. smoothing / kernel estimation smoothing / kernel estimation Quotation withheld because this record is not cleared for redistribution. Receiver operating characteristic Receiver operating characteristic Quotation withheld because this record is not cleared for redistribution. nonparametric resampling bootstrap nonparametric resampling bootstrap Quotation withheld because this record is not cleared for redistribution. 95% CI 95% CI Quotation withheld because this record is not cleared for redistribution. SEM SEM Quotation withheld because this record is not cleared for redistribution. Posterior parietal cortex encodes the running value of the accumulated evidence as a graded quantity, through a map that does not change over the course of the trial. Quotation withheld because this record is not cleared for redistribution. The frontal orienting fields instead encode mainly the sign of the accumulator, splitting activity into two clusters corresponding to the two available choices. Quotation withheld because this record is not cleared for redistribution. The FOF tuning curve was significantly steeper at the zero-crossing than the PPC tuning curve, quantifying the graded-versus-categorical distinction. Quotation withheld because this record is not cleared for redistribution. Clicks influenced FOF firing about 100 ms earlier than PPC firing, suggesting sensory input reaches FOF by a route that does not pass through PPC. Quotation withheld because this record is not cleared for redistribution. Silencing FOF across the whole trial produced a significant bias toward choices on the same side as the inactivation. Quotation withheld because this record is not cleared for redistribution. Of the four inactivation windows, only peri-choice inactivation produced a significant ipsilateral bias (10.6%). Quotation withheld because this record is not cleared for redistribution. With 250-ms windows only the peri-choice perturbation affected behaviour; perturbation 250 ms earlier had a smaller, non-significant effect. Quotation withheld because this record is not cleared for redistribution. Model comparison favored a bias applied after categorization over biases acting on the accumulation process itself. Quotation withheld because this record is not cleared for redistribution. The inactivation-induced bias was uncorrelated with how many clicks were played or how long the stimulus lasted, as expected if FOF is not part of the accumulator. Quotation withheld because this record is not cleared for redistribution. The authors conclude that evidence accumulation happens upstream of the FOF, challenging the view that prefrontal cortex implements accumulation. Quotation withheld because this record is not cleared for redistribution. THE CLASSICAL ANALYSIS MAKES THE TWO AREAS LOOK ALIKE and the tuning curve assay separates them Quotation withheld because this record is not cleared for redistribution. side-selective neurons in both PPC and FOF ramped upwards for preferred-direction stimuli with a slope proportional to the stimulus strength, and downwards for anti-preferred ones Quotation withheld because this record is not cleared for redistribution. parietal cortex encodes the accumulator as a graded value through a map stable across time Quotation withheld because this record is not cleared for redistribution. frontal cortex clusters by the sign of the accumulator, which is the choice currently favoured Quotation withheld because this record is not cleared for redistribution. the whole distinction reduces to the slope at the zero crossing, 0.058 against 0.158 Quotation withheld because this record is not cleared for redistribution. and it holds per neuron as a distribution shift, with the populations overlapping Quotation withheld because this record is not cleared for redistribution. the difference survived scaling the integration time constant and the sticky bound, varying response latencies, and fitting the behavioural model per rat or in aggregate Quotation withheld because this record is not cleared for redistribution. but it does not survive the end of the decision - during motor preparation the parietal code becomes categorical too Quotation withheld because this record is not cleared for redistribution. single clicks have a sustained effect on firing in both areas, slowly decaying only in frontal cortex Quotation withheld because this record is not cleared for redistribution. and the click latency is about 200 ms in parietal against 100 ms in frontal, read as a route that bypasses parietal cortex Quotation withheld because this record is not cleared for redistribution. THE STIMULUS IS DESIGNED SO THE ANALYSIS IS VALID: Poisson clicks are independent in time, which is what lets the residual isolate one click's effect Quotation withheld because this record is not cleared for redistribution. THE ENCODING CLAIM IS TURNED INTO A TIMING PREDICTION: if frontal cortex only converts evidence into a choice, silencing it should matter only at the end Quotation withheld because this record is not cleared for redistribution. full-trial silencing produced a 10.3 percent ipsilateral choice bias Quotation withheld because this record is not cleared for redistribution. of four 500 ms windows only the peri-choice one biased behaviour, and the early accumulation window was not different from zero Quotation withheld because this record is not cleared for redistribution. at 250 ms resolution the same held, with a window ending 250 ms earlier having no significant effect Quotation withheld because this record is not cleared for redistribution. and the peri-choice bias fully accounted for the full-trial bias, at ratios of 1.17 and 0.93 in the two groups Quotation withheld because this record is not cleared for redistribution. a post-categorization bias fitted the silencing data better than biases acting on the accumulation itself Quotation withheld because this record is not cleared for redistribution. THE FOUR RIVAL BIASES ARE SORTED BY WHEN THEY ACT - two throughout accumulation, two only at its end Quotation withheld because this record is not cleared for redistribution. each rival is a single added parameter on a model whose original bias term was one decision borderline about this quote: THE QUOTE ENDS ON A COLON: the paper goes on with the four sources of side bias it compared, each described in its own paragraph, too long to quote here; what this row states is in that list, not in the quoted words. Quotation withheld because this record is not cleared for redistribution. with control trials fitted first and each variant then fitted to the inactivation trials from that starting point Quotation withheld because this record is not cleared for redistribution. and the comparison made over 200 bootstrap resamples rather than once Quotation withheld because this record is not cleared for redistribution. no correlation was found between the induced bias and click counts or stimulus duration Quotation withheld because this record is not cleared for redistribution. ENCODING IS NOT NECESSITY: parietal cortex carries the graded code and is reported elsewhere not to be needed for the behaviour Quotation withheld because this record is not cleared for redistribution. the conclusion is that accumulation happens upstream of frontal cortex, against the prevailing view Quotation withheld because this record is not cleared for redistribution. 394 neurons were recorded in PPC of 4 rats and 397 in FOF of 6 rats, and all isolatable neurons encountered were recorded regardless of response properties Quotation withheld because this record is not cleared for redistribution. and 93 neurons in PPC (23%) and 128 in FOF (32%) differed significantly in pre-movement firing rate by the choice that followed Quotation withheld because this record is not cleared for redistribution. with an explicit rate floor and significance test as the criteria Quotation withheld because this record is not cleared for redistribution. preferred sides were mixed rather than uniformly contralateral, near 59 percent in both areas Quotation withheld because this record is not cleared for redistribution. THE ACCUMULATOR ESTIMATE USES THE CHOICE THE ANIMAL WENT ON TO MAKE, by backward propagation from the end of the trial Quotation withheld because this record is not cleared for redistribution. neurons were combined by weighting each by the inverse variance of its conditional distribution, so less noisy representations count for more Quotation withheld because this record is not cleared for redistribution. the model was fitted separately to each rat, and the analysis repeated with one aggregate fit Quotation withheld because this record is not cleared for redistribution. the two parameters probed are the ones measuring how far accumulation departs from optimality Quotation withheld because this record is not cleared for redistribution. and response lag was varied for both areas jointly, as a matrix rather than a single choice Quotation withheld because this record is not cleared for redistribution. graded and categorical encoding schemes were simulated with matched lags to predict what each would look like Quotation withheld because this record is not cleared for redistribution. model-based and model-free analyses were both used to confirm the rats were accumulating throughout the stimulus Quotation withheld because this record is not cleared for redistribution. INVESTIGATORS WERE NOT BLINDED to experimental groups during collection or analysis, stated outright Quotation withheld because this record is not cleared for redistribution. recording area was assigned randomly per rat and confirmed by histology Quotation withheld because this record is not cleared for redistribution. AND THE OTHER HALF OF THE DISCLOSURE: two rats were dropped for showing no full-trial effect, and the timing groups were not randomized Quotation withheld because this record is not cleared for redistribution. laser illumination occurred on a random quarter of trials interleaved with controls Quotation withheld because this record is not cleared for redistribution. Poisson Clicks task Poisson Clicks task Quotation withheld because this record is not cleared for redistribution. tuning curves tuning curves Quotation withheld because this record is not cleared for redistribution. click-triggered average click-triggered average Quotation withheld because this record is not cleared for redistribution. post-categorization bias post-categorization bias Quotation withheld because this record is not cleared for redistribution. accumulator shift accumulator shift Quotation withheld because this record is not cleared for redistribution. pre-movement period pre-movement period Quotation withheld because this record is not cleared for redistribution. tuning curve tuning curve Quotation withheld because this record is not cleared for redistribution. click-triggered average click-triggered average Quotation withheld because this record is not cleared for redistribution. post-categorization post-categorization Quotation withheld because this record is not cleared for redistribution. pre-movement period pre-movement period Quotation withheld because this record is not cleared for redistribution. Easy trials Easy trials Quotation withheld because this record is not cleared for redistribution. difficult trials difficult trials Quotation withheld because this record is not cleared for redistribution. trial difficulty trial difficulty Quotation withheld because this record is not cleared for redistribution. brain region brain region Quotation withheld because this record is not cleared for redistribution. ipsilateral choice bias ipsilateral choice bias Quotation withheld because this record is not cleared for redistribution. non-inactivation trials non-inactivation trials Quotation withheld because this record is not cleared for redistribution. total mean click rate held fixed at 40 Hz total mean click rate held fixed at 40 Hz Quotation withheld because this record is not cleared for redistribution. firing rate firing rate Quotation withheld because this record is not cleared for redistribution. accumulator accumulator Quotation withheld because this record is not cleared for redistribution. bias bias Quotation withheld because this record is not cleared for redistribution. latency latency Quotation withheld because this record is not cleared for redistribution. steepness steepness Quotation withheld because this record is not cleared for redistribution. 394 ... 397 — neurons recorded in PPC (from 4 rats) and in FOF (from 6 rats) Quotation withheld because this record is not cleared for redistribution. 23% ... 32% — proportion of recorded neurons in PPC and in FOF that were side-selective before movement onset Quotation withheld because this record is not cleared for redistribution. PPC: 0.058 ± 0.003, FOF: 0.158 ± 0.015 — slope of the tuning curve at the zero-crossing in each region; the roughly threefold steeper FOF slope is the paper's core quantitative contrast Quotation withheld because this record is not cleared for redistribution. 10.3 ± 3.0% — ipsilateral choice bias caused by full-trial unilateral FOF silencing Quotation withheld because this record is not cleared for redistribution. 5.4 ± 0.8% — ipsilateral bias from silencing only the final 250 ms of the click train Quotation withheld because this record is not cleared for redistribution. 59.1% ... 59.4% — share of side-selective neurons preferring the contralateral choice in PPC and FOF, i.e. only a mild contralateral bias in both regions Quotation withheld because this record is not cleared for redistribution. 0.058 ± 0.003, FOF: 0.158 ± 0.015 — slope at the zero crossing of the tuning curve - the single number the graded against categorical claim rests on Quotation withheld because this record is not cleared for redistribution. 10.3 ± 3.0% — ipsilateral choice bias from silencing the whole trial Quotation withheld because this record is not cleared for redistribution. 1.17 ± 0.45 and 0.93 ± 0.33 — peri-choice bias divided by full-trial bias in the two groups - the last window accounts for all of the effect Quotation withheld because this record is not cleared for redistribution. 394 neurons in PPC of 4 rats and 397 neurons in FOF of 6 rats — the recorded sample, taken without regard to response properties Quotation withheld because this record is not cleared for redistribution. ~200 ms — click latency in parietal cortex against about 100 ms in frontal cortex Quotation withheld because this record is not cleared for redistribution. 59.1% — share of parietal neurons preferring the contralateral side, against 59.4 percent in frontal cortex - selectivity is mixed 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/nature14066.