Behavior-dependent short-term assembly dynamics in the medial prefrontal cortex

Behavior-dependent short-term assembly dynamics in the medial prefrontal cortex Fujisawa Amarasingham Harrison Buzsáki

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.

Behavior-dependent short-term assembly dynamics in the medial prefrontal cortex 2008 Nature Neuroscience 10.1038/nn.2134 Fujisawa2008-behavior-dependent-short-term-assembly-dynam Fujisawa2008-behavior-dependent-short-term-assembly-dynam.txt, the project's staged copy 20c42d405275314ec2f0b86fb99c46c52b3050c8 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 short-term synaptic plasticity can be seen at work in an intact brain during behaviour Quotation withheld because this record is not cleared for redistribution. monosynaptic efficacy varies with position and trajectory in a working memory task, beyond what co-varying firing rates can explain Quotation withheld because this record is not cleared for redistribution. mechanism Quotation withheld because this record is not cleared for redistribution. behaviour-dependent short-term synaptic plasticity in prefrontal cortex Quotation withheld because this record is not cleared for redistribution. electrophysiology electrophysiology Quotation withheld because this record is not cleared for redistribution. mechanisms Short-term synaptic facilitation and depression modulate spike transmission, so sequentially discharging neurons form internally generated assemblies whose lifetimes are set by changing synaptic efficacies status: the paper's own words: We hypothesize that Quotation withheld because this record is not cleared for redistribution. mechanisms Supersynchronous presynaptic input triggers dendritic spikes in the target interneuron, producing supralinear summation of coincident inputs Quotation withheld because this record is not cleared for redistribution. mechanisms Neuronal reverberation: a receding assembly hands its representational content to a successor assembly, generating the ordered firing sequence Quotation withheld because this record is not cleared for redistribution. one presynaptic neuron's increased activity can reduce its control of an interneuron while another's increases it, on the same postsynaptic cell Quotation withheld because this record is not cleared for redistribution. that opposite pattern on one cell argues for synaptic mechanisms rather than passive membrane properties Quotation withheld because this record is not cleared for redistribution. coincident discharge of two or more presynaptic neurons within 5 ms raises the target's firing probability supralinearly Quotation withheld because this record is not cleared for redistribution. dendritic spikes triggered by supersynchronous inputs are offered as a mechanism for that supralinearity Quotation withheld because this record is not cleared for redistribution. sequences may reflect a receding assembly handing its content to the next, with lifetimes set by internal mechanisms including short-term plasticity Quotation withheld because this record is not cleared for redistribution. Only a small, possibly unrepresentative fraction of neurons and connections could be observed. what this limits: any population-level or connectivity claim: ONLY A SMALL PERCENTAGE OF NEURONS AND THEIR CONNECTIONS COULD BE MONITORED IN OUR STUDY Quotation withheld because this record is not cleared for redistribution. Synaptic connections are inferred indirectly from extracellular cross-correlograms, not measured. what this limits: every reported synaptic connection: MONOSYNAPTIC INTERACTIONS CAN ONLY BE INDIRECTLY INFERRED FROM AN EXTRACELLULAR SIGNAL - inferred, not measured Quotation withheld because this record is not cleared for redistribution. The cross-correlation method is insensitive to weak pyramidal-to-pyramidal interactions. what this limits: the absence of weak pyramidal-to-pyramidal effects specifically: THE CROSS-CORRELATION METHOD MAY NOT RELIABLY DETECT AND ANALYZE WEAK EXCITATORY INTERACTIONS AMONG PRINCIPAL CELLS - a null here is not evidence of absence Quotation withheld because this record is not cleared for redistribution. The jitter window size was a judgement call, which qualifies literal reading of the significance values. what this limits: the significance values themselves: CHOOSING THE SIZE OF THE JITTER WINDOW WAS A MATTER OF JUDGMENT, AND THIS SHOULD BE KEPT IN MIND IN THE LITERAL INTERPRETATION OF SIGNIFICANCE Quotation withheld because this record is not cleared for redistribution. The position-dependence analysis is example-based, without corrections for multiple comparisons across cells. what this limits: the position-dependence analysis: OUR APPROACH HERE IS EXAMPLE-BASED, AND WE DID NOT EXPLICITLY COMPUTE MULTIPLE-COMPARISONS CONTROLS ACROSS CELLS Quotation withheld because this record is not cleared for redistribution. The population firing-rate estimates need confirmation by more direct methods. what this limits: the population firing-rate estimates only: THE PRESENT ESTIMATES SHOULD BE CONFIRMED BY FUTURE STUDIES USING MORE DIRECT METHODS Quotation withheld because this record is not cleared for redistribution. only a small percentage of neurons could be monitored, most are silent or sparsely active, and the population estimates are stated to need more direct methods what this limits: the same scope as limitations[0], with the added reason that most neurons are silent or sparsely active and so never enter the sample Quotation withheld because this record is not cleared for redistribution. the cross-correlation method may not reliably detect weak excitatory interactions between principal cells, though it works for principal-to-interneuron connections what this limits: the same scope as limitations[2] Quotation withheld because this record is not cleared for redistribution. the observed principal-to-interneuron ratio is lower than anatomy predicts, attributed to the recording method or to silent principal cells what this limits: any use of the cell-type composition as an anatomical estimate: the observed ratio IS LOWER THAN would be expected from anatomy, and the authors attribute the difference to the recording method Quotation withheld because this record is not cleared for redistribution. the layer difference in trajectory selectivity is reported with the caveat that firing rate differences could influence it what this limits: the layer difference in trajectory selectivity: the authors report it with the caveat that firing-rate differences between layers could contribute to the effect Quotation withheld because this record is not cleared for redistribution. conditional inference by surrogate construction, with the null chosen to preserve every timescale but the one under test Quotation withheld because this record is not cleared for redistribution. thinning - deleting observations at random until a nuisance variable is flat - imported as a power control Quotation withheld because this record is not cleared for redistribution. spike trains were jittered at a chosen timescale and the co-firing recomputed, so that only excess co-firing at the FINE timescale survives eliminates: that a co-firing pair is explained by shared slow input rather than by a monosynaptic connection - co-firing at broad timescales is common without any direct connection · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. the timing of the interaction was compared against what a passive membrane could produce eliminates: that the temporal structure is a passive electrical artefact - the membrane time constant or a changing input conductance in the postsynaptic cell · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. supralinear supralinear Quotation withheld because this record is not cleared for redistribution. dendritic spikes dendritic spikes Quotation withheld because this record is not cleared for redistribution. variation in firing rates → detectability RULED OUT BY THE AUTHORS AS AN EXPLANATION: because spikes are uniformly distributed across maze positions, the observed differences in monosynaptic activity are LESS LIKELY TO BE DUE TO firing-rate variation acting on detectability. A control argument, not a relation that holds - and no edge type can say a relation FAILS. Quotation withheld because this record is not cleared for redistribution. MONOSYNAPTIC here is an inference to the simpler explanation, not an observation - the named rival is shared input producing co-firing without any synapse, and the criterion that separates them is TIMESCALE Quotation withheld because this record is not cleared for redistribution. the surrogate is built so that its null contains everything EXCEPT the timescale of interest: jittering by ±5 ms preserves firing rates at broader timescales, so what survives cannot be explained by rate variation at tens of milliseconds Quotation withheld because this record is not cleared for redistribution. functional synaptic efficacy is defined operationally as the magnitude of excess coincident spikes at short latency - it is a spike-count statistic, not a measured synaptic quantity Quotation withheld because this record is not cleared for redistribution. the independence assumption used by the standard shift-predictor approach is rejected here BY NAME, because broad-timescale effects alone can push short-latency co-occurrence above chance Quotation withheld because this record is not cleared for redistribution. the confound being controlled is DETECTABILITY rather than the effect: more spikes make a connection easier to detect, so a position-dependent rate would produce position-dependent detection with no change in efficacy Quotation withheld because this record is not cleared for redistribution. claim_type: existence a second reading DIFFERING from the recorded value. Recorded: mechanism — 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. method: eight-shank silicon probe recording with jitter-surrogate cross-correlogram analysis and a spike-thinning control for detection power 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: excess short-latency coincident spiking between a pyramidal cell and an interneuron, changing systematically across the maze a second reading DIFFERING from the recorded value. Recorded: behaviour-dependent short-term synaptic plasticity in prefrontal cortex — note: a second reading, verbatim-verified, disagreeing with the value on the record. Not applied - a person decides. Quotation withheld because this record is not cleared for redistribution. replicates → previous studies reporting that short-timescale cross-correlations between neuron pairs vary with behaviour (refs 14-16, 33) Quotation withheld because this record is not cleared for redistribution. contradicts → those earlier studies' assumption that behaviourally varying short-latency cross-correlation peaks reflect pyramidal-pyramidal interactions Quotation withheld because this record is not cleared for redistribution. extends → prior in vitro observations that excitatory inputs from different sources onto the same interneuron can be depressing or potentiating (refs 2, 5) Quotation withheld because this record is not cleared for redistribution. reinterprets → earlier reports of behaviour-dependent short-term cross-correlations, assumed to reflect pyramidal-to-pyramidal interactions Quotation withheld because this record is not cleared for redistribution. extends → in vitro evidence that pyramidal-to-interneuron synapses are plastic over tens to hundreds of milliseconds Quotation withheld because this record is not cleared for redistribution. builds-on → jitter and surrogate methods for separating fine-timescale from broad-timescale synchrony Quotation withheld because this record is not cleared for redistribution. critiques → the independence assumption underlying shift-predictor identification of monosynaptic connections differs in: THE AXIS IS THE INDEPENDENCE ASSUMPTION, STATED PRECISELY IN THE SPAN. 'Seeking more formal criteria for large scale analysis, a standard approach is Quotation withheld (that is, precisely, that the response of one neuron is conditionally exchangeable across trials, or shifts...).' The shift predictor is not wrong as arithmetic; it rests on an assumption the paper names in its formal form. Quotation withheld because this record is not cleared for redistribution. monosynaptic —operationalised-as→ cross-correlogram monosynaptic cross-correlogram strength of the relation: 1 · note: it can ONLY be inferred indirectly from an extracellular signal - the paper opens the section by saying so Quotation withheld because this record is not cleared for redistribution. facilitation —contrasted-with→ depression facilitation depression strength of the relation: 1 · note: both found, and elsewhere BOTH ON ONE POSTSYNAPTIC CELL from different presynaptic partners, which is what rules out passive membrane properties Quotation withheld because this record is not cleared for redistribution. cell assemblies —part-of→ mPFC cell assemblies mPFC strength of the relation: 0.8 · note: distributed across a large volume with no spatial clustering of similar tuning Quotation withheld because this record is not cleared for redistribution. resampling / surrogate data methods resampling / surrogate data methods Quotation withheld because this record is not cleared for redistribution. permutation test permutation test Quotation withheld because this record is not cleared for redistribution. spike-jitter resampling for cross-correlogram significance spike-jitter resampling for cross-correlogram significance Quotation withheld because this record is not cleared for redistribution. kernel density estimation kernel density estimation Quotation withheld because this record is not cleared for redistribution. multiple-comparison correction (family-wise error rate) multiple-comparison correction (family-wise error rate) Quotation withheld because this record is not cleared for redistribution. confidence interval confidence interval Quotation withheld because this record is not cleared for redistribution. t-test t-test Quotation withheld because this record is not cleared for redistribution. standard deviation standard deviation Quotation withheld because this record is not cleared for redistribution. randomization by thinning (statistical power control) randomization by thinning (statistical power control) Quotation withheld because this record is not cleared for redistribution. cross-correlation cross-correlation Quotation withheld because this record is not cleared for redistribution. jitter jitter Quotation withheld because this record is not cleared for redistribution. Clopper-Pearson confidence intervals Clopper-Pearson confidence intervals Quotation withheld because this record is not cleared for redistribution. On the seconds timescale, transiently active mPFC neurons formed sequences that reliably distinguished the rat's left and right maze trajectories. Quotation withheld because this record is not cleared for redistribution. On the millisecond timescale, the efficacy of putative monosynaptic connections was modulated by task phase even after firing-rate co-variation was statistically accounted for. Quotation withheld because this record is not cleared for redistribution. Roughly equal fractions of connected pairs showed firing-pattern-dependent synaptic depression and facilitation in the behaving animal. Quotation withheld because this record is not cleared for redistribution. Presynaptic spikes coinciding within 5 ms drove the postsynaptic interneuron supralinearly, more than the sum of their separate effects. Quotation withheld because this record is not cleared for redistribution. 16-18% of neurons also fired differentially in the central arm, where movement trajectories and head directions were apparently indistinguishable. Quotation withheld because this record is not cleared for redistribution. Putative connection probability fell rapidly with the distance between somata. Quotation withheld because this record is not cleared for redistribution. Information in mPFC appears sparsely encoded by cell assemblies spread over a large volume. Quotation withheld because this record is not cleared for redistribution. monosynaptic efficacy varied with position in the maze and with left versus right trajectory rather than staying constant Quotation withheld because this record is not cleared for redistribution. 67 of 343 pairs with significant excitatory connections showed identifiable position dependence Quotation withheld because this record is not cleared for redistribution. 495 of 62,408 directed cell pairs, 0.79 percent, had significant short-latency narrow peaks or troughs Quotation withheld because this record is not cleared for redistribution. connection probability fell rapidly with distance but connections were found up to 1,200 micrometres apart Quotation withheld because this record is not cleared for redistribution. THE INFERENCE IS TO THE SIMPLER EXPLANATION: the named rival is co-firing from shared input with no synapse, and the discriminator is timescale Quotation withheld because this record is not cleared for redistribution. THE STANDARD ASSUMPTION IS REJECTED BY NAME: under independence, broad-timescale effects alone can raise short-latency co-occurrence above chance Quotation withheld because this record is not cleared for redistribution. and an informal observation is given as further reason: the independence approach reliably yields ambiguous putative connections Quotation withheld because this record is not cleared for redistribution. each spike in each neuron was randomly and independently perturbed on a uniform interval of [-5,+5] ms, repeated independently 1,000 times to form the surrogate data sets Quotation withheld because this record is not cleared for redistribution. THE NULL IS BUILT TO CONTAIN EVERYTHING BUT THE TIMESCALE OF INTEREST: jittering preserves broader-timescale rates, so what survives cannot be rate variation at tens of milliseconds Quotation withheld because this record is not cleared for redistribution. THE CONFOUND IS DETECTABILITY: more spikes make a connection easier to detect, so a position-dependent rate could fake a position-dependent efficacy Quotation withheld because this record is not cleared for redistribution. SO DATA WERE DISCARDED TO REMOVE IT: spikes were deleted at random until rates were flat to within ten percent across maze segments, and the analysis rerun Quotation withheld because this record is not cleared for redistribution. the argument being that with spikes uniform across positions, a remaining position dependence is unlikely to be a power effect Quotation withheld because this record is not cleared for redistribution. and the conclusion survived thinning in several cases Quotation withheld because this record is not cleared for redistribution. in some pairs thinning was unnecessary because rates were already flat while coincidences were confined to part of the maze Quotation withheld because this record is not cleared for redistribution. A PRIOR LITERATURE REATTRIBUTED: earlier work read such peaks as pyramidal-to-pyramidal, and physiological classification makes them mostly pyramidal-to-interneuron Quotation withheld because this record is not cleared for redistribution. neurons fired differentially by trajectory already in the central arm, where movement and head direction were indistinguishable Quotation withheld because this record is not cleared for redistribution. the neuron already showed goal-specific elevated firing in the central arm, suggestive of goal representation Quotation withheld because this record is not cleared for redistribution. a lingering-odour confound was ruled out: stem-area differential firing showed no reliable relationship with nose-poke firing Quotation withheld because this record is not cleared for redistribution. and previous-trial reward position left only sparse traces in firing Quotation withheld because this record is not cleared for redistribution. environmental and motor differences alone cannot fully account for mPFC activity Quotation withheld because this record is not cleared for redistribution. individual cells fired at specific locations while the population as a whole fired uniformly over the apparatus Quotation withheld because this record is not cleared for redistribution. one presynaptic partner's activity reduced its control of an interneuron while another's increased it, on the same cell Quotation withheld because this record is not cleared for redistribution. which argues for synaptic mechanisms rather than passive membrane properties Quotation withheld because this record is not cleared for redistribution. coincident presynaptic discharge within 5 ms raised target firing probability supralinearly Quotation withheld because this record is not cleared for redistribution. THE SILENT MAJORITY ACKNOWLEDGED: only about 10 to 25 percent of neurons in the sampled volume were active enough to cluster, and the estimates are said to need more direct methods Quotation withheld because this record is not cleared for redistribution. the principal-to-interneuron ratio of 2.82 is lower than anatomy predicts, attributed to the method or to silent principal cells Quotation withheld because this record is not cleared for redistribution. 633 units were recorded from anterior cingulate and dorsal prelimbic cortex in four rats Quotation withheld because this record is not cleared for redistribution. rats performed at 92 percent correct at the time of recording Quotation withheld because this record is not cleared for redistribution. functional synaptic efficacy functional synaptic efficacy Quotation withheld because this record is not cleared for redistribution. spike train spike train Quotation withheld because this record is not cleared for redistribution. coincident-time offset spike coincident-time offset spike Quotation withheld because this record is not cleared for redistribution. surrogate data set surrogate data set Quotation withheld because this record is not cleared for redistribution. monosynaptic monosynaptic Quotation withheld because this record is not cleared for redistribution. jitter jitter Quotation withheld because this record is not cleared for redistribution. thinning thinning Quotation withheld because this record is not cleared for redistribution. position in the maze position in the maze Quotation withheld because this record is not cleared for redistribution. jittered data sets jittered data sets Quotation withheld because this record is not cleared for redistribution. firing rates firing rates Quotation withheld because this record is not cleared for redistribution. trajectory trajectory Quotation withheld because this record is not cleared for redistribution. 633 — Total well-isolated single units recorded from medial prefrontal cortex across four rats Quotation withheld because this record is not cleared for redistribution. 495 (0.79%) of 62,408 — Cell pairs showing significant short-latency cross-correlogram peaks or troughs, i.e. putative monosynaptic connections Quotation withheld because this record is not cleared for redistribution. 2.82 ± 0.51 — Physiologically identified ratio of putative principal cells to inhibitory interneurons Quotation withheld because this record is not cleared for redistribution. 67 of 343 — Excitatory monosynaptic pairs showing statistically identifiable position dependence of their interaction Quotation withheld because this record is not cleared for redistribution. 12.7% and 10.7% — Proportions of neuron pairs showing firing-pattern-dependent depression and facilitation respectively Quotation withheld because this record is not cleared for redistribution. 92% correct — Mean behavioural accuracy of the rats on the odour-place matching task at the time of recording Quotation withheld because this record is not cleared for redistribution. 40% versus 22% — Fraction of trajectory-selective neurons in the side arms in layer 5 versus layer 2/3 Quotation withheld because this record is not cleared for redistribution. 0.27 (62 cm) ± 0.17 (39 cm) — Mean standardized maze distance over which an individual neuron remained active (its firing 'lifetime') Quotation withheld because this record is not cleared for redistribution. 0.79% — share of directed cell pairs with a significant short-latency peak or trough Quotation withheld because this record is not cleared for redistribution. 67 of 343 — excitatory pairs showing position-dependent monosynaptic interaction Quotation withheld because this record is not cleared for redistribution. 1,200 μm — the greatest separation at which a connection was detected Quotation withheld because this record is not cleared for redistribution. 2.82 ± 0.51 — observed principal-to-interneuron ratio, against about 4 from anatomy Quotation withheld because this record is not cleared for redistribution. 92% correct — behavioural performance during recording 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/nn.2134.