The Origin of Adaptation in the Auditory Pathway of Locusts Is Specific to Cell Type and Function

The Origin of Adaptation in the Auditory Pathway of Locusts Is Specific to Cell Type and Function Hildebrandt Benda Hennig

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.

The Origin of Adaptation in the Auditory Pathway of Locusts Is Specific to Cell Type and Function 2009 The Journal of Neuroscience 10.1523/jneurosci.4800-08.2009 Hildebrandt2009-the-origin-of-adaptation-in-the-auditory-pat Hildebrandt2009-the-origin-of-adaptation-in-the-auditory-pat.txt, the project's staged copy 49a86f54a31f3c829954317a008f57355d90eaff 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 where in a layered sensory pathway the adaptation seen in one neuron actually arises Quotation withheld because this record is not cleared for redistribution. how to separate the contribution of intrinsic from presynaptic adaptation mechanisms in a single cell Quotation withheld because this record is not cleared for redistribution. adaptation mechanisms differ greatly between interneurons of one pathway and follow the cell's role in information processing Quotation withheld because this record is not cleared for redistribution. mechanism Quotation withheld because this record is not cleared for redistribution. spike frequency adaptation Quotation withheld because this record is not cleared for redistribution. electrophysiology electrophysiology Quotation withheld because this record is not cleared for redistribution. mechanisms In AN2 each spike triggers a slowly decaying afterhyperpolarizing current; during a stimulus these afterhyperpolarizations accumulate, progressively lowering excitability and thereby producing the adaptation. Quotation withheld because this record is not cleared for redistribution. mechanisms In AN2 the responsible conductance is inferred to be a spike-triggered AHP- or M-type current rather than a network effect. Quotation withheld because this record is not cleared for redistribution. mechanisms In BSN1 the hyperpolarization is not generated by the cell's own spikes but is imposed by presynaptic inhibitory input driving slow hyperpolarizing currents. Quotation withheld because this record is not cleared for redistribution. mechanisms A functional design rule: when a neuron's computation depends on the balance of incoming excitation and inhibition, adaptation must sit in the neuron itself rather than in its inputs, so that the balance is preserved. Quotation withheld because this record is not cleared for redistribution. current injection bypasses everything upstream of the cell, so the difference between the two stimulation modes is what the pathway contributes Quotation withheld because this record is not cleared for redistribution. in BSN1 the hyperpolarisation is not caused by the cell's own spiking, so it is probably imposed by inhibitory inputs Quotation withheld because this record is not cleared for redistribution. in AN2 long-lasting spike-triggered afterhyperpolarisations accumulate, pointing to an AHP- or M-type current Quotation withheld because this record is not cleared for redistribution. a strong enough intrinsic mechanism can overwrite the time course inherited from upstream, which is what AN2 does Quotation withheld because this record is not cleared for redistribution. a design principle stated as a conditional: if a neuron's computation depends on the ratio of inhibition to excitation, the neuron itself should adapt rather than its inputs Quotation withheld because this record is not cleared for redistribution. in BSN1 adaptation via one input would carry over to a later stimulus arriving on another, mixing temporal and directional information Quotation withheld because this record is not cleared for redistribution. TN1 conclusions hold only up to a limited response range because current injection could not reach the high firing rates evoked acoustically. what this limits: every TN1 conclusion, over a stated interval: OUR CONCLUSIONS FOR TN1 ARE STRICTLY VALID ONLY FOR ONSET RESPONSES UP TO 150 HZ Quotation withheld because this record is not cleared for redistribution. Linking cell-type-specific adaptation to temporal processing needs long-duration natural stimuli that were not used. what this limits: the link from cell-type-specific adaptation to temporal processing: EXPERIMENTS WITH LONG-DURATION NATURAL STIMULI ARE REQUIRED, and were not done here Quotation withheld because this record is not cleared for redistribution. Only the stimulus side eliciting the larger response was tested, so directional processing was not assessed. what this limits: any directional or binaural reading: BECAUSE WE TESTED STIMULI FROM THE SIDE THAT ELICITED A LARGER RESPONSE ... OUR PRESENT INVESTIGATION OF SFA WAS RESTRICTED TO TEMPORAL PROCESSING Quotation withheld because this record is not cleared for redistribution. AN2 could not be analysed at the higher reference onset response because only one cell reached it. what this limits: the AN2 results at the higher reference level: ONLY ONE OF THE CELLS UNDER STUDY RESPONDED WITH >140 HZ AT BOTH STIMULATION MODES Quotation withheld because this record is not cleared for redistribution. Large propagated errors prevented interpreting the relation between acoustic and current adaptation ratios in AN2. what this limits: the comparison between acoustic and current adaptation ratios: AT THESE LEVELS THE LARGE ERRORS RULE OUT AN INTERPRETATION OF THE RELATION BETWEEN THESE TWO Quotation withheld because this record is not cleared for redistribution. conclusions for TN1 hold only up to 150 Hz onset response, because current injection could not reach the responses that loud sound could what this limits: the same scope as limitations[0], with its cause: LARGE SOUND INTENSITIES ELICIT VERY STRONG RESPONSES NOT REACHED BY CURRENT STIMULATION - the instrument sets the bound Quotation withheld because this record is not cleared for redistribution. AN2 was not analysed at the higher onset response because only one cell reached it under both modes duplicate of: limitations[3] · what this limits: the same scope as limitations[3] Quotation withheld because this record is not cleared for redistribution. stimuli were tested from the side that elicited a larger response for each neuron, so the investigation of SFA was restricted to temporal processing what this limits: the same scope as limitations[2] - one side only, so temporal processing only Quotation withheld because this record is not cleared for redistribution. linking the cell-specific adaptation to temporal processing would need long-duration natural stimuli what this limits: the same scope as limitations[1] - long-duration natural stimuli are required and absent Quotation withheld because this record is not cleared for redistribution. a predicted correlation between adaptation under the two modes was absent in TN1, possibly masked by propagated errors what this limits: the mechanistic identification of the adaptation source in TN1: the prediction that CELLS EXHIBITING STRONGER SFA UNDER CURRENT STIMULATION SHOULD ALSO ADAPT MORE STRONGLY TO ACOUSTIC STIMULI was not borne out Quotation withheld because this record is not cleared for redistribution. the gain function and a high-pass cutoff predicted from a time constant, applied to spike trains Quotation withheld because this record is not cleared for redistribution. propagation of error by partial differentiation, carried through an interpolation into the reported ratios Quotation withheld because this record is not cleared for redistribution. spontaneous activity in the same three cells analysed BEFORE AND AFTER the current or acoustic stimulation eliminates: a drift in the cell's own baseline as the apparent adaptation - if spontaneous rate had changed by itself, the post-stimulus measurement would differ for a reason unrelated to the stimulus · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. BSN1 BSN1 Quotation withheld because this record is not cleared for redistribution. adaptation measured under CURRENT injection and under SOUND are two different quantities on purpose: the first excludes everything upstream of the electrode, the second includes it, and the paper's whole result is their difference Quotation withheld because this record is not cleared for redistribution. the adaptation ratio is defined at a FIXED onset response (70 or 140 Hz) reached by interpolation, because adaptation depends on intensity and the two stimulation modes cannot otherwise be compared about this quote: This quote ends on 'by' because what it introduces is a displayed formula that the staged extraction does not carry as readable text (the text continues: "r a ( I ref ) = f ss ( I ref )/ f on ref . Propa"). The quote is verbatim and the formula is in the paper; the quote was not extended, because extending it would put extraction debris inside a quotation. Quotation withheld because this record is not cleared for redistribution. an adaptation time constant enters the analysis only where a single exponential already fits with r>0.8, so the reported constants describe the subset of responses that are exponential at all Quotation withheld because this record is not cleared for redistribution. the three cell types are not a sample from a population but one identified neuron each per animal, named by morphology after staining Quotation withheld because this record is not cleared for redistribution. the three layers are feedforward with no within-layer connections, which is what licenses reading the three cells as three processing levels 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: intracellular recording from the same identified cell under both acoustic and injected-current stimulation, with exponential fits and gain functions 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: spike frequency adaptation, the decrease in response strength during prolonged stimulation a second reading DIFFERING from the recorded value. Recorded: spike frequency adaptation — 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. replicates → Weschke and Ronacher (2008) Quotation withheld because this record is not cleared for redistribution. builds-on → Benda (2002) Quotation withheld because this record is not cleared for redistribution. replicates → Prinz and Ronacher (2002); Wohlgemuth and Ronacher (2007); Weschke and Ronacher (2008) Quotation withheld because this record is not cleared for redistribution. extends → the finding that spike frequency adaptation is distributed across levels in vertebrate visual, somatosensory and auditory systems Quotation withheld because this record is not cleared for redistribution. confirms → Weschke and Ronacher's high-pass characteristic for BSN1 with a cutoff between 10 and 20 Hz Quotation withheld because this record is not cleared for redistribution. builds-on → Benda's measurement that locust receptor cells adapt to about half their onset response Quotation withheld because this record is not cleared for redistribution. qualifies → studies that infer adaptation mechanisms from current stimulation in vitro or sensory stimulation in vivo alone differs in: THE AXIS IS THE PREPARATION. 'In the latter case, intrinsic properties of isolated cells or small circuits are isolated, Quotation withheld.' Two literatures measure the same named phenomenon with methods that cannot see the same mechanisms - the qualification is about what each preparation can possibly show. Quotation withheld because this record is not cleared for redistribution. current stimulation —contrasted-with→ acoustic stimulation current stimulation acoustic stimulation differs in: THE AXIS IS WHETHER PERIPHERAL ADAPTATION IS INCLUDED IN WHAT IS MEASURED. 'During acoustic stimulation, Quotation withheld.' The same cellular process is measured both ways; one route carries everything upstream of the cell with it. · strength of the relation: 1 · note: THE WHOLE DESIGN IN ONE SENTENCE: sound convolves the cell's own adaptation with every process upstream of it; injected current does not. The difference IS the pathway's contribution Quotation withheld because this record is not cleared for redistribution. TN1 —part-of→ receptor TN1 receptor strength of the relation: 0.8 · note: TN1 receives monosynaptic receptor input and behaves like a receptor, which is why the null model applies to it most cleanly Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span describes a three-layer network; TN1 part-of 'receptor' is loose. Quotation withheld because this record is not cleared for redistribution. AN2 —causes→ afterhyperpolarization AN2 afterhyperpolarization strength of the relation: 1 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but AN2 REVEALED long-lasting afterhyperpolarizations, an intrinsic property; causes is loose. Quotation withheld because this record is not cleared for redistribution. time constant —operationalised-as→ high-pass time constant high-pass strength of the relation: 1 · note: A CLOSED-FORM PREDICTION: the cutoff is one over two-pi-tau, so a constant measured in one experiment predicts a curve measured in another Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span says the high-pass characteristic is predicted by SFA time constants; operationalised-as is loose. Quotation withheld because this record is not cleared for redistribution. low-pass —contrasted-with→ high-pass low-pass high-pass strength of the relation: 1 · note: CRITERION: the two cutoffs are set by different quantities - mean spike frequency and adaptation time constant - and two TN1 curves at different intensities move one while leaving the other alone Quotation withheld because this record is not cleared for redistribution. RAM —operationalised-as→ gain function RAM gain function strength of the relation: 1 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but RAM stimuli were PRESENTED to calculate gain functions; operationalised-as is loose. Quotation withheld because this record is not cleared for redistribution. BSN1 —contrasted-with→ AN2 BSN1 AN2 strength of the relation: 1 · note: CRITERION: they look alike under sound and come apart under current - which is exactly the comparison the design exists to make Quotation withheld because this record is not cleared for redistribution. TN1 —contrasted-with→ BSN1 TN1 BSN1 strength of the relation: 1 · note: and the reverse: alike under current, different under sound Quotation withheld because this record is not cleared for redistribution. encoder adaptation —part-of→ receptor encoder adaptation receptor strength of the relation: 1 · note: and the TN1 time constants match the receptor range, which is why no third mechanism is needed for TN1 Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span says SFA arises from encoder adaptation in receptor cells; part-of is loose. Quotation withheld because this record is not cleared for redistribution. Wilcoxon rank-sum test Wilcoxon rank-sum test Quotation withheld because this record is not cleared for redistribution. p-value p-value Quotation withheld because this record is not cleared for redistribution. correlation coefficient correlation coefficient Quotation withheld because this record is not cleared for redistribution. linear regression linear regression Quotation withheld because this record is not cleared for redistribution. nonlinear (exponential) curve fitting with model comparison nonlinear (exponential) curve fitting with model comparison Quotation withheld because this record is not cleared for redistribution. goodness-of-fit threshold for inclusion goodness-of-fit threshold for inclusion 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. standard error of the mean standard error of the mean Quotation withheld because this record is not cleared for redistribution. propagation of error propagation of error Quotation withheld because this record is not cleared for redistribution. linear interpolation linear interpolation Quotation withheld because this record is not cleared for redistribution. spectral (gain function) analysis spectral (gain function) analysis Quotation withheld because this record is not cleared for redistribution. Wilcoxon ranks sum test Wilcoxon ranks sum test Quotation withheld because this record is not cleared for redistribution. regression coefficient regression coefficient Quotation withheld because this record is not cleared for redistribution. Three identified interneurons of the locust auditory pathway show three qualitatively different distributions of adaptation mechanisms: mixed intrinsic/presynaptic in TN1, inhibition-driven in BSN1, and intrinsic spike-triggered in AN2. Quotation withheld because this record is not cleared for redistribution. The time course of adaptation is cell-type specific, differing significantly between the three neuron types. Quotation withheld because this record is not cleared for redistribution. Spike frequency adaptation is not confined to one station but operates at every level of the sensory pathway. Quotation withheld because this record is not cleared for redistribution. Adaptation time constants predicted the cells' signal transmission properties. Quotation withheld because this record is not cleared for redistribution. All three neurons displayed bandpass characteristics. Quotation withheld because this record is not cleared for redistribution. Which adaptation mechanism a neuron uses tracks the computation it performs, so the origin of adaptation is a function of the cell's processing role. Quotation withheld because this record is not cleared for redistribution. three distinct distributions of adaptation mechanisms emerged across three cell types of one pathway Quotation withheld because this record is not cleared for redistribution. AN ARITHMETIC NULL MODEL, NOT A DIRECTION: linear composition predicts the acoustic ratio equals 0.5 times the current ratio, and the data were compared against that line Quotation withheld because this record is not cleared for redistribution. in TN1 the magnitude of SFA to acoustic stimuli was qualitatively in agreement with a linear combination of SFA in receptors and adaptation mechanisms within TN1 Quotation withheld because this record is not cleared for redistribution. A PREDICTED CORRELATION THAT FAILED: cells adapting more under current should adapt more under sound, and TN1 showed r = 0.09, p = 0.81 Quotation withheld because this record is not cleared for redistribution. BSN1 at 140 Hz adapted far more than the null model allows, implying processes the model does not contain Quotation withheld because this record is not cleared for redistribution. AN ANALYSIS DECLINED AND SAID SO: AN2 was not analysed at the higher onset response because only one cell reached it under both modes Quotation withheld because this record is not cleared for redistribution. time constants were similar across modes in TN1 and in AN2, and differed in BSN1 where sound gave 18.7 ms against 49.8 ms for current Quotation withheld because this record is not cleared for redistribution. TN1 and BSN1 showed very similar adaptation characteristics under current stimulation, this pattern changed dramatically in BSN1 under acoustic stimulation, and in AN2 the two modes gave very similar results Quotation withheld because this record is not cleared for redistribution. THE TWO-SOURCE LOGIC: an intrinsic current would hyperpolarise after either stimulus, but only sound lets another neuron inhibit the cell Quotation withheld because this record is not cleared for redistribution. in BSN1, post-stimulus hyperpolarisation tracked acoustic response strength (r = -0.93) while current stimulation had almost no effect Quotation withheld because this record is not cleared for redistribution. in AN2 hyperpolarisation tracked response strength under both modes, around r = -0.92 Quotation withheld because this record is not cleared for redistribution. CONFOUND NAMED: an absent hyperpolarisation could mean the electrode was far from the spike initiating zone rather than that no current exists Quotation withheld because this record is not cleared for redistribution. CONFOUND ANSWERED BY CHANGING THE OBSERVABLE: spontaneous spike counts were used instead, since the recording site cannot affect them Quotation withheld because this record is not cleared for redistribution. the spike-count control gave the same answer, and the two analyses agree Quotation withheld because this record is not cleared for redistribution. A NUMERICAL CEILING ON THEIR OWN CLAIM: conclusions for TN1 are stated to be strictly valid only up to 150 Hz onset response Quotation withheld because this record is not cleared for redistribution. afterhyperpolarisation decay differed between all three cell types: 5.6 ms in TN1, 24.6 ms in BSN1, 118.3 ms in AN2 Quotation withheld because this record is not cleared for redistribution. TWO INDEPENDENT MEASUREMENTS AGREEING: in AN2 the afterhyperpolarisation decay resembles the adaptation time constant Quotation withheld because this record is not cleared for redistribution. the spike shapes explain the offset results: TN1 decays too fast to see, BSN1 has almost none so its hyperpolarisation must be extrinsic, AN2 accumulates Quotation withheld because this record is not cleared for redistribution. A CLOSED-FORM PREDICTION FROM AN INDEPENDENTLY MEASURED CONSTANT: the high-pass cutoff should be one over two-pi-tau Quotation withheld because this record is not cleared for redistribution. AND IT HELD FOR ALL THREE: in TN1 the gain decreased for frequencies <3 Hz where the prediction lied at 2.6 Hz, in BSN1 a higher cutoff at 10 Hz was observed with 8.5 predicted, and in AN2 a cutoff between 1 and 2 Hz with 1.1 predicted Quotation withheld because this record is not cleared for redistribution. the low-pass cutoff is set instead by mean spike frequency, separable in one cell at two intensities Quotation withheld because this record is not cleared for redistribution. BSN1 and AN2 end up alike under sound while TN1 stays tonic and locks less selectively Quotation withheld because this record is not cleared for redistribution. and the BSN1-AN2 similarity DEVELOPS over seconds - it is absent in the first 100 ms Quotation withheld because this record is not cleared for redistribution. in AN2 adaptation is stronger but slower than in receptors, so an intrinsic mechanism can overwrite an inherited time course Quotation withheld because this record is not cleared for redistribution. 18 cells in all: nine TN1, five BSN1, four AN2 Quotation withheld because this record is not cleared for redistribution. MODEL ORDER CHECKED: double exponential fits did not significantly improve on single exponentials Quotation withheld because this record is not cleared for redistribution. ERRORS PROPAGATED EXPLICITLY through the interpolation by partial differentiation Quotation withheld because this record is not cleared for redistribution. spike frequency adaptation (SFA) spike frequency adaptation (SFA) Quotation withheld because this record is not cleared for redistribution. rate coefficient rate coefficient about this quote: THE QUOTE ENDS ON A COLON WHERE THE PAPER DISPLAYS AN EQUATION: the text extraction does not carry the formula, so what this row states rests on that equation, not on the quoted words. Quotation withheld because this record is not cleared for redistribution. SFA SFA Quotation withheld because this record is not cleared for redistribution. adaptation ratio adaptation ratio about this quote: This quote ends on 'by' because what it introduces is a displayed formula that the staged extraction does not carry as readable text (the text continues: "r a ( I ref ) = f ss ( I ref )/ f on ref . Propa"). The quote is verbatim and the formula is in the paper; the quote was not extended, because extending it would put extraction debris inside a quotation. Quotation withheld because this record is not cleared for redistribution. high-pass high-pass Quotation withheld because this record is not cleared for redistribution. stimulation mode stimulation mode Quotation withheld because this record is not cleared for redistribution. stimulus intensity stimulus intensity Quotation withheld because this record is not cleared for redistribution. cell type cell type Quotation withheld because this record is not cleared for redistribution. hyperpolarization hyperpolarization Quotation withheld because this record is not cleared for redistribution. gain function gain function Quotation withheld because this record is not cleared for redistribution. onset spike frequency onset spike frequency Quotation withheld because this record is not cleared for redistribution. adaptation ratio adaptation ratio Quotation withheld because this record is not cleared for redistribution. time constant time constant Quotation withheld because this record is not cleared for redistribution. spontaneous activity spontaneous activity Quotation withheld because this record is not cleared for redistribution. 18 — number of adult locusts used in the experiments Quotation withheld because this record is not cleared for redistribution. TN1, n = 9; BSN1, n = 5; AN2, n = 4 — number of recorded cells of each of the three identified types (18 cells in total) Quotation withheld because this record is not cleared for redistribution. 60.0 ± 14.9 ms; 49.3 ± 21.6 ms; p = 0.37 — TN1 adaptation time constants under acoustic and current stimulation, statistically indistinguishable Quotation withheld because this record is not cleared for redistribution. 49.8 ± 14.8 ms vs 18.7 ± 10.9 ms (p < 0.01) — BSN1 adaptation time constants for current versus acoustic stimulation, which do differ Quotation withheld because this record is not cleared for redistribution. 143.6 ± 38.2 ms; 123.8 ± 35.0 ms; p = 0.68 — AN2 adaptation time constants under acoustic and current stimulation, both slow and not different Quotation withheld because this record is not cleared for redistribution. 5.6 ± 0.7 ms / 24.6 ± 10.7 ms / 118.3 ± 34.7 ms — decay time constants of the afterhyperpolarization in TN1, BSN1 and AN2, all significantly different Quotation withheld because this record is not cleared for redistribution. 2.9 ± 4.3% vs 14.8 ± 4.3% and 12.8 ± 1.5% — relative afterhyperpolarization magnitude in BSN1 compared with TN1 and AN2 Quotation withheld because this record is not cleared for redistribution. r = 0.09, p = 0.81 — absence of correlation in TN1 between adaptation ratios under acoustic and under current stimulation Quotation withheld because this record is not cleared for redistribution. current: r = 0.04, p = 0.92, acoustic: r = −0.93, p = 0.0001 — BSN1 correlation between response strength and post-stimulus membrane potential, present only for acoustic stimulation Quotation withheld because this record is not cleared for redistribution. <3 Hz observed vs 2.6 Hz predicted (τ = 60.0 ms) — TN1 high-pass cutoff of the gain function versus the cutoff predicted from its adaptation time constant Quotation withheld because this record is not cleared for redistribution. 10 Hz observed vs 8.5 Hz predicted (mean τ = 18.7 ms) — BSN1 high-pass cutoff versus the prediction from its faster adaptation time constant Quotation withheld because this record is not cleared for redistribution. ∼0.5 — BSN1 adaptation ratio under current stimulation, constant across intensities Quotation withheld because this record is not cleared for redistribution. 0.5 × r a cu — the linear-composition prediction the acoustic adaptation ratio is tested against Quotation withheld because this record is not cleared for redistribution. r = 0.09, p = 0.81 — the predicted cross-mode correlation in TN1, absent Quotation withheld because this record is not cleared for redistribution. 18.7 ± 10.9 ms against 49.8 ± 14.8 ms — BSN1 adaptation time constant, acoustic against current Quotation withheld because this record is not cleared for redistribution. 5.6, 24.6 and 118.3 ms — afterhyperpolarisation decay in TN1, BSN1 and AN2 Quotation withheld because this record is not cleared for redistribution. 150 Hz — the onset response above which the TN1 conclusions are stated not to hold Quotation withheld because this record is not cleared for redistribution. n = 9, 5 and 4 — cells of each type 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.4800-08.2009.