Reactivation of hippocampal ensemble memories during sleep

Reactivation of hippocampal ensemble memories during sleep Wilson McNaughton

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.

Reactivation of hippocampal ensemble memories during sleep 1994 Science 10.1126/science.8036517 Wilson1994-reactivation-of-hippocampal-ensemble-memorie Wilson1994-reactivation-of-hippocampal-ensemble-memorie.txt, the project's staged copy 6c9e78d414d32da2e01a8e4c850cdc8359374a19 not-redistributable neuroscience

the paper's own words

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

0 spans located by searching the stored text for the quote's exact words, allowing any whitespace between them the activity correlations between pairs of hippocampal CA1 cells, measured across sleep before behaviour, behaviour, and sleep after it Quotation withheld because this record is not cleared for redistribution. whether the interactions among hippocampal cells change in a way SPECIFIC to the experience just had, rather than changing in general Quotation withheld because this record is not cleared for redistribution. the first demonstration that cells coactive during behaviour become more correlated in the sleep that follows, while cells that were active and NOT coactive do not Quotation withheld because this record is not cleared for redistribution. existence Quotation withheld because this record is not cleared for redistribution. hippocampal ensemble reactivation during sleep Quotation withheld because this record is not cleared for redistribution. electrophysiology electrophysiology Quotation withheld because this record is not cleared for redistribution. mechanisms Waking experience drives associative synaptic modification, which leaves cells that were coactive during behaviour more likely to be reactivated together in subsequent slow-wave sleep Quotation withheld because this record is not cleared for redistribution. mechanisms Event memory is first stored by rapid synaptic modification in the hippocampus, and is then replayed during slow-wave sleep (when further hippocampal modification is suppressed) so that the information is gradually transferred to neocortex Quotation withheld because this record is not cleared for redistribution. mechanisms Initial storage of the event occurs through rapid synaptic modification primarily within the hippocampus itself Quotation withheld because this record is not cleared for redistribution. an autoassociative network in which RECURRENT EXCITATION reproduces a stored activity pattern from very little input - so background noise alone can push the network into a stored state, seen as a population burst Quotation withheld because this record is not cleared for redistribution. random background noise occasionally lands the network in a correlated state, and the resulting population bursts ARE the ripples - the paper's account of why the effect appears in discrete events rather than continuously Quotation withheld because this record is not cleared for redistribution. storage is fast synaptic modification in the hippocampus during waking; sleep is the playback, and the transfer target is neocortex Quotation withheld because this record is not cleared for redistribution. The increased correlation implies associative synaptic change somewhere in the system, but not necessarily inside the hippocampus. Quotation withheld because this record is not cleared for redistribution. Neocortical or subcortical changes are an alternative source of the increased correlations. Quotation withheld because this record is not cleared for redistribution. THE PAPER REFUSES TO LOCATE ITS OWN MECHANISM. The increased correlation is called consistent with associative synaptic modification `at some stage in the system, although not necessarily in the hippocampus` - the structure it recorded from Quotation withheld because this record is not cleared for redistribution. a PAIR-WISE correlation cannot show that the activity is organised into a coherent representation; the paper says the prior single-cell result was suggestive and could not settle it, which is the same limit its own measure has Quotation withheld because this record is not cleared for redistribution. the population encoding any one event is expected to be small, so the whole design depends on recording enough cells for the pair count to grow as the square Quotation withheld because this record is not cleared for redistribution. cell pairs that were ACTIVE during running but NOT COACTIVE - non-overlapping place fields - are measured through the same PRE and POST sleep sessions as the coactive pairs eliminates: any explanation that would raise correlations across the board: arousal, electrode drift, general post-behaviour excitability, or a recording artefact of the sleep session itself · outcome: held Quotation withheld because this record is not cleared for redistribution. the RUN phase is checked first for the effect the sorting assumes - that overlapping place fields do produce correlated firing and non-overlapping ones do not eliminates: a sort that does not separate what it claims to, which would make the POST comparison meaningless in either direction · outcome: held Quotation withheld because this record is not cleared for redistribution. the correlation is computed separately DURING ripples and in the intervals between them eliminates: an effect spread evenly through sleep, which would not support the proposed burst-based transfer mechanism · outcome: absent-not-in-staged-text Quotation withheld because this record is not cleared for redistribution. whether the correlation increase can be localised to the hippocampus at all eliminates: nothing - the paper states the modification may be at some other stage in the system · outcome: absent-declared-unreported Quotation withheld because this record is not cleared for redistribution. place fields place fields Quotation withheld because this record is not cleared for redistribution. cross-correlations cross-correlations Quotation withheld because this record is not cleared for redistribution. ripples ripples Quotation withheld because this record is not cleared for redistribution. autoassociative memory autoassociative memory Quotation withheld because this record is not cleared for redistribution. synaptic modification synaptic modification Quotation withheld because this record is not cleared for redistribution. reactivation - AS MEASURED HERE: an increase in the short-latency cross-correlation of a cell PAIR from PRE to POST sleep. Not a replayed sequence, not a retrieved memory, not a hippocampal event specifically - the instrument sees a pair statistic and the title names a memory. The paper marks that gap twice rather than closing it Quotation withheld because this record is not cleared for redistribution. memory, AS THE PAPER USES IT HERE: `the structure of distributed representations of the visited locations`, retained across sleep. The definition is spatial and population-level, and it is what the overlap-sorted pairs operationalise Quotation withheld because this record is not cleared for redistribution. ripple - a synchronised intermittent burst during slow-wave sleep, identified with the EEG sharp wave. Used here as the WINDOW in which the correlation effect is concentrated, and as the network signature of a stored pattern being re-entered Quotation withheld because this record is not cleared for redistribution. the ripple/inter-ripple split - the paper's own test of where in sleep the effect lives. The design is stated plainly; its RESULT is the one sentence in this record that OCR damage puts out of reach, and the control below says so rather than borrowing a verdict from its neighbour Quotation withheld because this record is not cleared for redistribution. claim_type: existence a second reading reaching the same value. Recorded: existence — note: a second reading reaching the same value from a different sentence. Not applied - a person decides. Quotation withheld because this record is not cleared for redistribution. method: simultaneous recording of 50 to 100 CA1 cells, short-latency spike-train cross-correlation computed per pair over each of three phases, pairs sorted by how much their place fields overlap 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: experience-specific reactivation - coactivity acquired during waking behaviour re-expressed in hippocampal circuits during subsequent slow-wave sleep a second reading DIFFERING from the recorded value. Recorded: hippocampal ensemble reactivation during sleep — 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 → previous theoretical models of autoassociative memory (ref. 6) Quotation withheld because this record is not cleared for redistribution. extends → the large-ensemble recording method itself, used here because the number of observable pair-wise interactions grows as the square of the cells recorded - the argument that makes an experience-specific effect measurable at all Quotation withheld because this record is not cleared for redistribution. critiques → the indirect synaptic-efficacy evidence for hippocampal learning, on the ground that no change had been shown that was SPECIFIC to a given experience Quotation withheld because this record is not cleared for redistribution. supports → autoassociative network models in which recurrent excitation reconstructs a stored pattern from little input Quotation withheld because this record is not cleared for redistribution. ANOVA ANOVA Quotation withheld because this record is not cleared for redistribution. cross-correlation cross-correlation Quotation withheld because this record is not cleared for redistribution. mean mean Quotation withheld because this record is not cleared for redistribution. cross-correlations cross-correlations Quotation withheld because this record is not cleared for redistribution. correlation correlation Quotation withheld because this record is not cleared for redistribution. Hippocampal CA1 cells that fired together during a spatial behavioural task showed increased tendency to fire together in the sleep that followed, relative to the sleep that preceded the task. Quotation withheld because this record is not cleared for redistribution. The increase was specific to coactive pairs: cells silent during behaviour, and active cells whose place fields did not overlap, showed no such increase. Quotation withheld because this record is not cleared for redistribution. Coactive (overlapping place field) pairs rose significantly above their pre-behaviour correlation level, whereas pairs of spatially firing cells with non-overlapping fields did not. Quotation withheld because this record is not cleared for redistribution. The reactivation effect decayed across the post-behaviour sleep session with a time constant of about 12 minutes. Quotation withheld because this record is not cleared for redistribution. The sleep correlations were concentrated in the synchronized population bursts (ripples) rather than in the intervals between them. Quotation withheld because this record is not cleared for redistribution. During running, cells with overlapping place fields were strongly positively correlated while non-overlapping pairs were not, confirming the grouping used for the sleep comparison. Quotation withheld because this record is not cleared for redistribution. cells coactive during behaviour show a SIGNIFICANT INCREASE in correlation from PRE to POST sleep, and pairs that fired spatially but not in the same region do not - the effect is specific to shared experience, not to having been active cells coactive during behaviour show a SIGNIFICANT INCREASE in correlation from PRE to POST sleep, and pairs that fired spatially but not in the same region do not - the effect is specific to shared experience, not to having been active Quotation withheld because this record is not cleared for redistribution. THE PAPER DECLINES TO PLACE ITS OWN MECHANISM: the increase is consistent with associative synaptic modification at some stage in the system, ALTHOUGH NOT NECESSARILY IN THE HIPPOCAMPUS THE PAPER DECLINES TO PLACE ITS OWN MECHANISM: the increase is consistent with associative synaptic modification at some stage in the system, ALTHOUGH NOT NECESSARILY IN THE HIPPOCAMPUS Quotation withheld because this record is not cleared for redistribution. the sorting variable is validated before it is used - during RUN, overlapping place fields are highly positively correlated and non-overlapping ones are not the sorting variable is validated before it is used - during RUN, overlapping place fields are highly positively correlated and non-overlapping ones are not Quotation withheld because this record is not cleared for redistribution. large simultaneously recorded ensembles were studied because the number of observable pair interactions grows as the SQUARE of the sample size large simultaneously recorded ensembles were studied because the number of observable pair interactions grows as the SQUARE of the sample size Quotation withheld because this record is not cleared for redistribution. waking activity retains the structure of distributed representations of the visited locations across sleep - stated as the paper's conclusion waking activity retains the structure of distributed representations of the visited locations across sleep - stated as the paper's conclusion Quotation withheld because this record is not cleared for redistribution. intermittent, synchronized bursts in which cells discharge during slow-wave sleep, associated with EEG sharp-wave activity ripples Quotation withheld because this record is not cleared for redistribution. a model in which recurrent excitation reproduces a previously stored activity pattern even when the input contains very little of the relevant information autoassociative memory Quotation withheld because this record is not cleared for redistribution. amount of overlap amount of overlap Quotation withheld because this record is not cleared for redistribution. cross-correlation of spike trains cross-correlation of spike trains Quotation withheld because this record is not cleared for redistribution. significant spatial firing significant spatial firing Quotation withheld because this record is not cleared for redistribution. place fields place fields Quotation withheld because this record is not cleared for redistribution. overlap overlap Quotation withheld because this record is not cleared for redistribution. approximately 12 min — time constant of the decay of the reactivated correlation during post-behaviour sleep Quotation withheld because this record is not cleared for redistribution. rat 1: 49 cells; rat 2: 74 cells; rat 3: 69 cells — number of simultaneously recorded hippocampal pyramidal cells in each of the three rats Quotation withheld because this record is not cleared for redistribution. 50 to 100 — number of single CA1 cells monitored simultaneously in each recording session Quotation withheld because this record is not cleared for redistribution. NON 437 and OVR 60; NON 719 and OVR 95; NON 941 and OVR 147 — numbers of non-overlapping versus overlapping cell pairs contributing to the histograms for rats 1, 2 and 3 Quotation withheld because this record is not cleared for redistribution. P < 0.05 — significance level at which the post-run correlation increase for coactive pairs was established by ANOVA Quotation withheld because this record is not cleared for redistribution. approximately 20 min — duration of each of the PRE, RUN and POST recording periods Quotation withheld because this record is not cleared for redistribution. minutes - the approximate duration of each of the three phases, PRE, RUN and POST 20 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/10.1126/science.8036517.