Generation of Spike Latency Tuning by Thalamocortical Circuits in Auditory Cortex

Generation of Spike Latency Tuning by Thalamocortical Circuits in Auditory Cortex Zhou Mesik Sun Liang Xiao Tao Zhang

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.

Generation of Spike Latency Tuning by Thalamocortical Circuits in Auditory Cortex 2012 The Journal of Neuroscience 10.1523/jneurosci.1384-12.2012 Zhou2012-generation-of-spike-latency-tuning-by-thalam Zhou2012-generation-of-spike-latency-tuning-by-thalam.txt, the project's staged copy ca38d6d6ac15d921692b01cb6871d7d390186454 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 spike latency tuning in layer 4 pyramidal neurons - whether the sharpness of the code is relayed from the thalamus or built locally Quotation withheld because this record is not cleared for redistribution. whether latency tuning survives the convergence of many thalamic inputs onto one cortical cell, or has to be recreated after that convergence destroys it Quotation withheld because this record is not cleared for redistribution. the input latency tuning is nearly flat and the spike latency tuning is three to four times sharper, so the code is generated locally by a push-and-pull of sharply tuned intracortical excitation and broadly tuned inhibition acting on the integration time Quotation withheld because this record is not cleared for redistribution. mechanism Quotation withheld because this record is not cleared for redistribution. spike latency tuning for sound frequency Quotation withheld because this record is not cleared for redistribution. electrophysiology electrophysiology Quotation withheld because this record is not cleared for redistribution. mechanisms Differential frequency tuning of the excitatory and inhibitory input strengths creates two opposing effects on integration time (excitation shortens it at the best frequency, inhibition lengthens it away from it), which together stretch the spike latency tuning curve. Quotation withheld because this record is not cleared for redistribution. mechanisms Because inhibition is more broadly tuned than excitation, the excitation/inhibition ratio falls at the receptive field edges, which further prolongs the already long integration time there. Quotation withheld because this record is not cleared for redistribution. mechanisms Convergence of many thalamocortical axons with different characteristic frequencies onto one cortical neuron degrades the latency tuning carried by each axon, because the summed input onset is set by the single fastest input. Quotation withheld because this record is not cleared for redistribution. sharply tuned intracortical excitation shortens the integration time most at the optimal frequency, while broadly tuned inhibition prolongs it most at off-optimal frequencies - two antagonistic effects that together triple the dynamic range Quotation withheld because this record is not cleared for redistribution. and because inhibition is more broadly tuned than excitation, the excitation-to-inhibition ratio is worst at the receptive field periphery, where the integration time is already longest Quotation withheld because this record is not cleared for redistribution. with the onset of this second component is 2.6 ms after the initial onset of the excitatory response, while the first evoked spike is at 4.2 ms Quotation withheld because this record is not cleared for redistribution. the aggregate thalamic input is flat because it is set by the shortest latency among many convergent inputs, so the sharpness carried by each individual afferent is destroyed by the summation Quotation withheld because this record is not cleared for redistribution. QX-314 in the pipette prevented recording real spikes from the same voltage-clamped cells, so spike responses had to be modeled. what this limits: any spike-response claim about the recorded cells: DUE TO THE USE OF QX-314 IN THE INTRAPIPETTE SOLUTION TO IMPROVE THE RECORDING QUALITY, WE WERE UNABLE TO EXAMINE THE BONA FIDE SPIKE RESPONSES OF THE SAME NEURON Quotation withheld because this record is not cleared for redistribution. Inhibitory responses were not recorded in the cortical-silencing experiments. what this limits: the cortical-silencing experiments: TO SIMPLIFY THE EXPERIMENT AND IMPROVE THE SUCCESS RATE, INHIBITORY RESPONSES WERE NOT RECORDED Quotation withheld because this record is not cleared for redistribution. Non-specific effects of the pharmacological silencing mixture cannot be entirely ruled out. what this limits: the silencing results: NONSPECIFIC EFFECTS OF THE MIXTURE APPLICATION ... CANNOT BE TOTALLY EXCLUDED Quotation withheld because this record is not cleared for redistribution. The identity of the early-spiking neurons proposed to supply the fast intracortical excitation is unknown. what this limits: the proposed relay population: THE IDENTITIES OF THESE PRESUMPTIVE EXCITATORY RELAY NEURONS ARE ELUSIVE Quotation withheld because this record is not cleared for redistribution. Whether the cortex actually uses the proposed population-onset time reference is not established. what this limits: the functional claim: WHETHER THIS STRATEGY IS INDEED IMPLEMENTED BY THE CORTICAL CIRCUITS REMAINS AN OPEN QUESTION Quotation withheld because this record is not cleared for redistribution. the drug that made the recording possible made the target unmeasurable - QX-314 blocks the spikes of the very cell whose spikes are the output being explained, so they were derived from a model instead what this limits: every spike response reported: DUE TO THE USE OF QX-314 ... WE WERE UNABLE TO EXAMINE THE BONA FIDE SPIKE RESPONSES OF THE SAME NEURON - the drug enabling the recording blocks the spikes, so they are derived rather than observed Quotation withheld because this record is not cleared for redistribution. inhibition was not recorded at all in the silencing experiments, and the missing frequency range was replaced by a 55% threshold taken from a different set of recordings what this limits: the silencing experiments specifically: TO SIMPLIFY THE EXPERIMENT AND IMPROVE THE SUCCESS RATE, INHIBITORY RESPONSES WERE NOT RECORDED Quotation withheld because this record is not cleared for redistribution. the cells the mechanism requires have never been recorded from - their identities are called elusive, and the recording methods are stated to be biased against them what this limits: the population size the mechanism depends on: THEIR NUMBER COULD BE UNDERESTIMATED SINCE THEY ARE REPORTED TO BE RESISTANT TO GOLGI IMPREGNATION AND DIFFICULT TO SAMPLE WITH INTRACELLULAR RECORDINGS - the absence is a property of the methods Quotation withheld because this record is not cleared for redistribution. nonspecific effects of the silencing mixture cannot be totally excluded, and the argument for specificity is that the changes appear in both the temporal and the frequency domain duplicate of: limitations[2] · what this limits: the same scope as limitations[2] - specificity argued rather than demonstrated Quotation withheld because this record is not cleared for redistribution. and whether the brain actually uses a population onset as its time reference remains an open question what this limits: the same scope as limitations[4] - whether the brain uses a population onset as its time reference is open Quotation withheld because this record is not cleared for redistribution. first spike latency as a code for stimulus identity, imported as the property to be explained Quotation withheld because this record is not cleared for redistribution. the integrate-and-fire neuron fed with measured conductances, used to obtain the spikes the recording method destroyed Quotation withheld because this record is not cleared for redistribution. a four-layer thalamocortical network with background activity, used to ask whether input latency tuning is necessary at all Quotation withheld because this record is not cleared for redistribution. the requirement for an internal time reference, since the organism has no access to stimulus onset Quotation withheld because this record is not cleared for redistribution. spike latency measured with excitatory input only, and again with both excitatory and inhibitory input present eliminates: the contribution of inhibition to latency tuning - the excitation-only condition is the same cell without the thing being tested · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. A PROPERTY EVERY ELEMENT HAS, THAT THEIR SUM DOES NOT: individual thalamic neurons are sharply latency-tuned, but the aggregate input to one cortical cell is FLAT, because the sum is set exclusively by the SHORTEST latency among the convergent inputs Quotation withheld because this record is not cleared for redistribution. and the network model reproduces exactly that - clear tuning in each thalamic neuron, flat tuning in the aggregate, sharp tuning again in the output Quotation withheld because this record is not cleared for redistribution. so the sharpness is NOT INHERITED but reconstructed: the input dynamic range is about 1 ms and the output about 4 ms Quotation withheld because this record is not cleared for redistribution. THE SAME INPUT, THE SAME STRENGTH AT THE OPTIMAL FREQUENCY, AND OPPOSITE EFFECTS ON THE CODE: sharply tuned intracortical excitation INCREASED the dynamic range where cotuned excitation DECREASED it - the tuning width decides, not the amplitude Quotation withheld because this record is not cleared for redistribution. PUSH AND PULL: sharply tuned excitation shortens the integration time at the optimal frequency, broadly tuned inhibition prolongs it off-optimum - and an initial tuning of input latency is then NOT ABSOLUTELY REQUIRED Quotation withheld because this record is not cleared for redistribution. THE DRUG THAT MADE THE RECORDING POSSIBLE MADE THE TARGET UNMEASURABLE: QX-314 was used to improve recording quality, and the bona fide spikes of the same neuron could then not be examined - the output had to be DERIVED Quotation withheld because this record is not cleared for redistribution. and the derivation is validated the right way - sequential cell-attached and whole-cell recording FROM THE SAME NEURON, with the derived and recorded receptive fields and latency tunings compared directly Quotation withheld because this record is not cleared for redistribution. A MISSING MEASUREMENT REPLACED BY A NUMBER FROM ANOTHER EXPERIMENT: inhibition was not recorded in the silencing condition, so the spiking frequency range was set by thresholding the excitatory tuning curve at 55%, a ratio measured elsewhere Quotation withheld because this record is not cleared for redistribution. THE CELL THE MECHANISM REQUIRES IS THE ONE THE METHOD CANNOT SEE: spiny stellate cells are resistant to Golgi impregnation and hard to sample intracellularly, and every current recording method shows SAMPLING BIAS TOWARD LARGE PYRAMIDAL NEURONS Quotation withheld because this record is not cleared for redistribution. with the honest consequence stated - the identities of the presumptive excitatory relay neurons are ELUSIVE, and the proposal rests on reasoning rather than on having recorded them Quotation withheld because this record is not cleared for redistribution. THE ZERO POINT OF THE MEASURE BELONGS TO THE EXPERIMENTER, NOT TO THE ORGANISM: the brain has NO ACCESS to the external stimulus onset, so latency is well defined for the observer and undefined for the system being studied Quotation withheld because this record is not cleared for redistribution. and the proposed substitutes are all internal - oscillations, local field potentials, population onsets - with whether any is actually used left an open question Quotation withheld because this record is not cleared for redistribution. THREE ALTERNATIVES ELIMINATED BY TIMING ALONE: feedback and recurrent inputs arrive after the first spike, a recurrent component would appear at about 4 ms and was not evident, and the small jitter rules out a random early-firing subgroup Quotation withheld because this record is not cleared for redistribution. and gamma oscillations are excluded by arithmetic - a 2 ms delay is difficult to explain by a 30 to 80 Hz rhythm Quotation withheld because this record is not cleared for redistribution. THE TWO CODES NEED DIFFERENT AMOUNTS OF DATA: a smooth latency tuning curve is established in a SINGLE TRIAL, where the spike-rate curve relies on averaging multiple trials Quotation withheld because this record is not cleared for redistribution. and the case against the latency code is answered rather than dismissed - noise is thought more destructive to latency than to rate coding, but pooling helps both, and jitter relative to other neurons may be smaller than jitter relative to the stimulus 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: spikes derived from recorded excitatory and inhibitory conductances with an integrate-and-fire model, validated against cell-attached spikes recorded sequentially from the same neuron 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: first spike latency tuned for sound frequency - a timing code whose dynamic range is about 1 ms at the input and about 4 ms at the output a second reading DIFFERING from the recorded value. Recorded: spike latency tuning for sound frequency — 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. builds-on → the cortical silencing method of Liu et al. (2007) Quotation withheld because this record is not cleared for redistribution. replicates → Liu et al. (2007), on flattened excitatory frequency tuning after cortical silencing Quotation withheld because this record is not cleared for redistribution. replicates → Wu et al. (2008) and Sun et al. (2010), on inhibition being more broadly tuned than excitation differs in: THE AXIS IS RELATIVE TUNING BREADTH, AND IT IS WHAT IS BEING REPLICATED. 'THE TUNING CURVE OF INHIBITORY INPUTS APPEARED BROADER THAN THAT OF EXCITATORY INPUTS TO THE SAME CELL, CONSISTENT WITH PREVIOUS REPORTS.' A 'replicates' usually records that two papers agree; naming the property says WHAT AGREES - and 'to the same cell' is the control that makes the comparison mean anything. Quotation withheld because this record is not cleared for redistribution. extends → the thalamocortical network model of de la Rocha et al. (2008) Quotation withheld because this record is not cleared for redistribution. contradicts → the assumption that spike latency tuning is relayed unchanged along the ascending pathway differs in: THE AXIS IS WHERE IN THE CIRCUIT THE LATENCY IS MEASURED. 'the onset latency of their aggregate thalamic input exhibited nearly flat tuning for sound frequency, whereas their spike latency tuning was much sharper.' Relay and local generation are not rival readings of one measurement; they are claims about different stages, and the paper measures both. Quotation withheld because this record is not cleared for redistribution. supports → the feedforward-inhibition account of cortical response sharpening, extended from spike rate to spike timing Quotation withheld because this record is not cleared for redistribution. explains-away → the apparent need for latency-tuned thalamic input, since the model generates output tuning from flat input tuning Quotation withheld because this record is not cleared for redistribution. distinguishes-from → feedback, recurrent and oscillatory sources of the second excitatory component, each excluded on timing grounds differs in: THE AXIS IS ARRIVAL TIME RELATIVE TO THE FIRST SPIKE. 'In principle, Quotation withheld.' Three candidate sources are excluded by a single argument about WHEN they could act - not by measuring them, and not by denying they exist. An elimination that costs no experiment and holds only for the quantity being explained, which is first-spike latency. Quotation withheld because this record is not cleared for redistribution. critiques → the objection that noise destroys latency codes more than rate codes, on the grounds that pooling helps both and the relevant reference is other neurons differs in: THE AXIS IS WHAT EACH CODE IS ASSUMED TO BE READ BY. 'Quotation withheld together, while with latency codes, single spikes ar[e considered].' The comparison is unfair by construction: one code is given a population and the other a single spike, so the noise result follows from the assumption rather than from the codes. Quotation withheld because this record is not cleared for redistribution. predicts → that spiny stellate neurons, if they can be recorded in vivo, will be the postulated fast excitatory relay Quotation withheld because this record is not cleared for redistribution. inhibition —causes→ integration time inhibition integration time strength of the relation: 1 · note: broadly tuned inhibition PROLONGS the integration time preferentially at off-optimal frequencies Quotation withheld because this record is not cleared for redistribution. excitation —causes→ integration time excitation integration time strength of the relation: 1 · note: sharply tuned intracortical excitation SHORTENS it selectively at the optimal frequency Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span credits broadly tuned INHIBITION with prolonging integration time; excitation's role is not stated here. Quotation withheld because this record is not cleared for redistribution. integration time —causes→ dynamic range integration time dynamic range status: the paper's own words: likely · strength of the relation: 1 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the push-pull mechanisms expand the dynamic range; integration time as the cause is loose. Quotation withheld because this record is not cleared for redistribution. inhibition —contrasted-with→ excitation inhibition excitation strength of the relation: 1 Quotation withheld because this record is not cleared for redistribution. inhibition —causes→ receptive field inhibition receptive field strength of the relation: 1 · note: broader inhibitory tuning imposes larger E/I ratios at receptive field PERIPHERIES, where the integration time is already long Audit 2026-09-28: DOUBTFUL - read against its span and kept, but broader inhibitory tuning sharpens spike tuning; causes 'receptive field' is loose. Quotation withheld because this record is not cleared for redistribution. recurrent —distinct-from→ feedforward recurrent feedforward strength of the relation: 1 · note: THREE ALTERNATIVES ELIMINATED BY TIMING: feedback and recurrent inputs arrive after the first spike, a recurrent component would show at about 4 ms and WAS NOT EVIDENT, and the small jitter rules out a random early-firing subgroup Audit 2026-09-28: DOUBTFUL - read against its span and kept, but feedforward, feedback and recurrent are listed as alternative sources; distinct-from is loose. Quotation withheld because this record is not cleared for redistribution. tuning —contrasted-with→ spike tuning spike differs in: THE AXIS IS HOW MANY TRIALS THE TUNING NEEDS TO EXIST. 'Quotation withheld.' A code available on one trial and a code that exists only in the average are different claims about what a downstream neuron could use - and the averaging that makes the second visible is the analyst's, not the brain's. Same paper whose aggregation operator is already in this corpus for deciding the meaning of a sum. · strength of the relation: 1 · note: THE TWO CODES NEED DIFFERENT AMOUNTS OF DATA: the latency curve is established in a SINGLE TRIAL, the spike-rate curve relies on averaging Audit 2026-09-28: DOUBTFUL - read against its span and kept, but latency tuning emerges on a single trial while rate tuning needs averaging; contrasted-with is loose. Quotation withheld because this record is not cleared for redistribution. precision —part-of→ coding precision coding strength of the relation: 1 · note: two prerequisites stated - precise spike timing per presentation, and a broad dynamic range; submillisecond precision with a 4 ms range is judged sufficient Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but precise spike timing is one of 'two prerequisites' for latency coding; part-of is loose. Quotation withheld because this record is not cleared for redistribution. dynamic range —part-of→ coding dynamic range coding strength of the relation: 1 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but a broad dynamic range is one of 'two prerequisites' for latency coding; part-of is loose. Quotation withheld because this record is not cleared for redistribution. t-test t-test Quotation withheld because this record is not cleared for redistribution. t-test t-test Quotation withheld because this record is not cleared for redistribution. regression regression Quotation withheld because this record is not cleared for redistribution. coefficient of determination coefficient of determination Quotation withheld because this record is not cleared for redistribution. mean and standard deviation mean and standard deviation Quotation withheld because this record is not cleared for redistribution. spline regression smoothing spline regression smoothing Quotation withheld because this record is not cleared for redistribution. paired t test paired t test Quotation withheld because this record is not cleared for redistribution. mean ± SD mean ± SD Quotation withheld because this record is not cleared for redistribution. The onset latency of the aggregate thalamic input to layer 4 neurons is almost untuned for sound frequency, while the same cells' spike latency is sharply tuned with a much wider dynamic range. Quotation withheld because this record is not cleared for redistribution. Broadly tuned intracortical inhibition lengthens the integration time for spike generation at off-optimal frequencies, whereas sharply tuned intracortical excitation shortens it at the optimal frequency. Quotation withheld because this record is not cleared for redistribution. The dynamic range of spike latency is three times that of synaptic input latency. Quotation withheld because this record is not cleared for redistribution. outputs of individual thalamic neurons exhibited a clear latency tuning, and the aggregate input to each pyramidal neuron had a flat latency tuning Quotation withheld because this record is not cleared for redistribution. Cortical pyramidal neurons produced sharp spike latency tuning even when their input latency tuning was flat. Quotation withheld because this record is not cleared for redistribution. the onset of this second component is 2.6 ms after the initial onset of the excitatory response, while that of the first evoked spike is 4.2 ms Quotation withheld because this record is not cleared for redistribution. First spike latency was tuned much like spike rate, with a similar preferred frequency. Quotation withheld because this record is not cleared for redistribution. Adding inhibition to the derived responses lengthened spike latency mostly at the tuning-curve tails, widening the dynamic range and sharpening latency tuning. Quotation withheld because this record is not cleared for redistribution. the input latency tuning is nearly flat and the spike latency tuning much sharper, so the code is NOT simply inherited from the thalamus Quotation withheld because this record is not cleared for redistribution. broadly tuned intracortical inhibition prolongs the integration time at off-optimal frequencies while sharply tuned intracortical excitation shortens it at the optimal frequency Quotation withheld because this record is not cleared for redistribution. with the stated strategy being the conversion of information spatially coded in synaptic strength into a temporal representation Quotation withheld because this record is not cleared for redistribution. THE SUM LOSES WHAT EVERY ELEMENT HAS: the summed input latency is determined exclusively by the SHORTEST latency among many convergent inputs, so a sharp output tuning has to be recreated from a flat input tuning Quotation withheld because this record is not cleared for redistribution. and the network model reproduces it - clear latency tuning in each thalamic neuron, flat tuning in the aggregate, sharp tuning in the output anyway Quotation withheld because this record is not cleared for redistribution. the dynamic range is about 1 ms for the input and about 4 ms for the spike output Quotation withheld because this record is not cleared for redistribution. measured stepwise: 1.5 ms of input latency, 2.0 ms from thalamocortical input alone, 3.1 ms once intracortical excitation is included Quotation withheld because this record is not cleared for redistribution. the onset of this second component is 2.6 ms after the initial onset of the excitatory response, while the first evoked spike is at 4.2 ms Quotation withheld because this record is not cleared for redistribution. PUSH AND PULL, with the consequence that an initial tuning of input latency is NOT ABSOLUTELY REQUIRED for the creation of spike latency tuning Quotation withheld because this record is not cleared for redistribution. inhibition is more broadly tuned than excitation, so the E/I ratio is worst at receptive field peripheries where the integration time is already longest Quotation withheld because this record is not cleared for redistribution. cotuned and sharply tuned intracortical excitation shortened integration time equally at the optimal frequency, but the sharply tuned kind increased the dynamic range of spike latency while the cotuned kind decreased it Quotation withheld because this record is not cleared for redistribution. the derived spikes were validated against cell-attached spikes recorded sequentially FROM THE SAME NEURON, on receptive field, best frequency and dynamic range Quotation withheld because this record is not cleared for redistribution. THE DRUG THAT IMPROVED THE RECORDING REMOVED THE OBSERVATION: with QX-314 in the pipette the bona fide spikes of the same neuron could not be examined, so they were derived from a model Quotation withheld because this record is not cleared for redistribution. A MISSING MEASUREMENT REPLACED BY A NUMBER FROM ANOTHER EXPERIMENT: inhibition was not recorded in the silencing condition and the spiking frequency range was set by thresholding at 55% Quotation withheld because this record is not cleared for redistribution. spiny stellate cells, the candidate relay, resist Golgi impregnation and are hard to sample intracellularly, and current patch-recording methods are biased toward large pyramidal neurons Quotation withheld because this record is not cleared for redistribution. with the identities of the presumptive excitatory relay neurons called ELUSIVE and the proposal resting on reasoning rather than on recordings Quotation withheld because this record is not cleared for redistribution. THE ZERO POINT OF THE MEASURE BELONGS TO THE EXPERIMENTER: the brain has no access to the external stimulus onset, so an internally generated time reference is necessary Quotation withheld because this record is not cleared for redistribution. three candidate internal references are named - oscillations, local field potentials, population onsets - and whether any is used is left an open question Quotation withheld because this record is not cleared for redistribution. THREE ALTERNATIVES ELIMINATED BY TIMING ALONE: feedback and recurrent inputs arrive too late, the expected recurrent component was not evident, and the small jitter excludes a random early-firing subgroup Quotation withheld because this record is not cleared for redistribution. and gamma oscillations are excluded by arithmetic - a 2 ms delay is difficult to explain by a 30 to 80 Hz rhythm Quotation withheld because this record is not cleared for redistribution. nonspecific effects of the silencing mixture cannot be totally excluded, and the specificity argument is that the changes appear in both the temporal and the frequency domain Quotation withheld because this record is not cleared for redistribution. THE TWO CODES NEED DIFFERENT AMOUNTS OF DATA: the latency tuning curve is established in a single trial, the spike-rate curve relies on averaging multiple trials Quotation withheld because this record is not cleared for redistribution. two prerequisites for a latency code are stated - precise spike timing per presentation and a broad dynamic range - and submillisecond precision with a 4 ms range is judged sufficient WHOSE RESULT THIS IS: Previous studies have demonstrated - THE SUBJECT OF THIS RESULT IS AN INDEFINITE STUDY, NOT THIS PAPER. 'Previous studies reported', 'studies in guinea pig showed', 'recent studies have reported' - somebody else's work under a common noun rather than a name, with a citation in the sentence and no first-person marker. The quote is verbatim and the claim is true; only the attribution is wrong. FOUND BY MEASURING MY OWN CUE'S RECALL: the first sweep required a NAMED author as the subject ('Sperling (1960) demonstrated') and found 17. Reading eleven citation-bearing findings on 'Plakke2016' by hand showed four the cue could not see - 'A study examining recognition found', 'studies of typical human subjects have demonstrated', 'A follow up study found' - so the named cue's 17 was a floor and not a count. A CUE'S FALSE POSITIVES TELL YOU NOTHING ABOUT ITS FALSE NEGATIVES, and the only way to find these was to read a record's whole layer. Quotation withheld because this record is not cleared for redistribution. the standing objection that noise hurts latency codes more than rate codes is answered: pooling helps both, and jitter relative to other neurons may be lower than relative to the stimulus Quotation withheld because this record is not cleared for redistribution. with 20 Hz background activity the latency tuning is largely preserved, but the jitter increases especially at off-optimal frequencies Quotation withheld because this record is not cleared for redistribution. first spike latency (FSL) first spike latency (FSL) Quotation withheld because this record is not cleared for redistribution. input latency input latency Quotation withheld because this record is not cleared for redistribution. integration time integration time Quotation withheld because this record is not cleared for redistribution. integration time integration time Quotation withheld because this record is not cleared for redistribution. dynamic range dynamic range Quotation withheld because this record is not cleared for redistribution. reference reference Quotation withheld because this record is not cleared for redistribution. feedforward feedforward Quotation withheld because this record is not cleared for redistribution. tone-evoked excitatory and inhibitory synaptic conductances tone-evoked excitatory and inhibitory synaptic conductances Quotation withheld because this record is not cleared for redistribution. integration time integration time Quotation withheld because this record is not cleared for redistribution. dynamic range dynamic range Quotation withheld because this record is not cleared for redistribution. jitter jitter Quotation withheld because this record is not cleared for redistribution. frequency frequency Quotation withheld because this record is not cleared for redistribution. a factor of three — how much broader the spike-latency dynamic range is than the input-latency dynamic range Quotation withheld because this record is not cleared for redistribution. 1.5 ± 0.5 ms — dynamic range of the isolated thalamocortical input latency across frequency Quotation withheld because this record is not cleared for redistribution. 3.1 ± 0.8 ms — dynamic range of derived spike latency once intracortical excitation was included Quotation withheld because this record is not cleared for redistribution. 14.34 ± 0.42 ms — average first spike latency in the intracellularly recorded layer 4 neurons Quotation withheld because this record is not cleared for redistribution. 2.6 ± 0.7 ms — delay of the second (intracortical) excitatory component after the onset of excitation Quotation withheld because this record is not cleared for redistribution. 4.2 ± 0.3 ms — delay of the first evoked spike after the onset of excitation Quotation withheld because this record is not cleared for redistribution. 93 — number of middle-layer neurons whose first spike latencies were sampled with cell-attached recording Quotation withheld because this record is not cleared for redistribution. ∼55% — fraction of the excitatory response range over which the cell actually spikes, used to threshold the tuning curve Quotation withheld because this record is not cleared for redistribution. 800 — number of model cells in each of the four layers of the thalamocortical network model Quotation withheld because this record is not cleared for redistribution. ∼1 ms ... ∼4 ms — the dynamic range of latency tuning within the frequency range for spiking response - for the input, and increased for the spike output Quotation withheld because this record is not cleared for redistribution. 1.5 ± 0.5 ms ... 2.0 ± 0.8 ms ... 3.1 ± 0.8 ms — the same range measured stepwise as each input component is added back, so the mechanism is attributed by subtraction rather than asserted Quotation withheld because this record is not cleared for redistribution. 2.6 ± 0.7 ms ... 4.2 ± 0.3 ms — the intracortical component arrives before the first spike, which is the entire license for saying it can modulate spike timing Quotation withheld because this record is not cleared for redistribution. 55% — a ratio measured in current-clamp recordings and used as a threshold in a different experiment where inhibition was never recorded Quotation withheld because this record is not cleared for redistribution. 20 Hz — background firing set deliberately ABOVE the typical level, so preserved tuning is a conservative result 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/10.1523/jneurosci.1384-12.2012.