Internally Generated Reactivation of Single Neurons in Human Hippocampus During Free Recall

Internally Generated Reactivation of Single Neurons in Human Hippocampus During Free Recall Gelbard-Sagiv Mukamel Harel Malach Fried

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.

Internally Generated Reactivation of Single Neurons in Human Hippocampus During Free Recall 2008 Science (New York, N.Y.) 10.1126/science.1164685 Gelbard-Sagiv2008-internally-generated-reactivation-of-single Gelbard-Sagiv2008-internally-generated-reactivation-of-single.txt, the project's staged copy 0e5f884ff74cba1e5e42cfa3b67adb729193d60c 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 single neurons in human hippocampus and entorhinal cortex, recorded while patients view cinematic episodes and again while they freely recall them Quotation withheld because this record is not cleared for redistribution. how MTL neurons respond to episodes presented as film clips, and what they do later when those episodes spontaneously come to mind with no external stimulus present Quotation withheld because this record is not cleared for redistribution. shows that the same single neurons that fired during viewing fire again seconds before the verbal report of free recall, giving spontaneous human recollection a single-neuron correlate Quotation withheld because this record is not cleared for redistribution. existence Quotation withheld because this record is not cleared for redistribution. single-neuron reactivation during free recall Quotation withheld because this record is not cleared for redistribution. electrophysiology electrophysiology Quotation withheld because this record is not cleared for redistribution. mechanisms Free recall arises when the same medial temporal neurons that encoded an episode are internally reactivated in the absence of external cues, producing the felt spontaneous emergence of a memory status: the paper's own words: We suggest that Quotation withheld because this record is not cleared for redistribution. temporal precedence carries the causal weight: the neurons rise above baseline about 1500 ms before the verbal report, so the report cannot be what drives the firing Quotation withheld because this record is not cleared for redistribution. the confound is ruled out by signature mismatch rather than by exclusion: epileptic activity is synchronous across neighbouring cells, whereas the observed responses are sparse and isolated among non-responsive neighbours Quotation withheld because this record is not cleared for redistribution. an anatomical dissociation inside the same experiment serves as the control condition: medial frontal units are selective during viewing but lose that selectivity at recall Quotation withheld because this record is not cleared for redistribution. sparseness plus the size of the untested stimulus space is used to infer, without testing, that each neuron must also respond to many episodes never presented Quotation withheld because this record is not cleared for redistribution. the recall signal is claimed to be readable on one trial rather than only after averaging, which is what lets a single spontaneous recall event be treated as an observation Quotation withheld because this record is not cleared for redistribution. the recall alignment is to the onset of the patient's own verbal report, so the experiment's time zero is supplied by the subject rather than by the experimenter Quotation withheld because this record is not cleared for redistribution. Findings come from epilepsy patients, so pathology cannot be ruled out with certainty. what this limits: any generalisation from these recordings to the healthy brain: ALTHOUGH THESE RESULTS SHOULD BE VIEWED WITH CAUTION, SUCH INTERPRETATION IS UNLIKELY FOR VARIOUS REASONS - argued against, not excluded Quotation withheld because this record is not cleared for redistribution. Only a tiny fraction of possible episodes was tested, so the apparent selectivity of the neurons is an artefact of the limited stimulus set. what this limits: the selectivity claim itself: IT IS CLEAR, EVEN BY A SIMPLE STATISTICAL REASONING, THAT THESE NEURONS MUST DISPLAY SELECTIVE RESPONSES TO MULTIPLE OTHER CLIPS not tested here Quotation withheld because this record is not cleared for redistribution. Which feature of a clip a cell actually responds to cannot be determined. what this limits: any claim about WHICH feature drove the response: IT IS ALSO IMPORTANT TO EXERCISE CAUTION IN CLAIMS AS TO WHAT EXACT ASPECT OF THE CLIPS THE CELL RESPONDED TO Quotation withheld because this record is not cleared for redistribution. Regions that might carry top-down recall signals were not sampled. what this limits: the claim that recall signals originate where they were recorded: Quotation withheld Quotation withheld because this record is not cleared for redistribution. Whether the multiple clips a neuron responds to share an abstract association is unresolved. what this limits: any abstract-association reading: WHETHER MULTIPLE CLIPS TO WHICH A NEURON RESPONDS MAY BE RELATED BY SOME ABSTRACT ASSOCIATION RULE IS NOT CLEAR AT PRESENT Quotation withheld because this record is not cleared for redistribution. which brain regions were sampled was decided by clinical need, not by the hypothesis — the coverage of the study is a by-product of surgical planning what this limits: the anatomical coverage of the whole study: Quotation withheld Quotation withheld because this record is not cleared for redistribution. the origin of the recall signal may lie in regions the electrodes never reached what this limits: the same scope as limitations[3] - the origin of the recall signal may be outside the sampled regions Quotation withheld because this record is not cleared for redistribution. the subjects all have intractable epilepsy, so a pathology explanation cannot be excluded outright, only argued against duplicate of: limitations[0] · what this limits: the same scope as limitations[0] - the subjects all have intractable epilepsy Quotation withheld because this record is not cleared for redistribution. what feature of a clip a cell is actually responding to is undetermined what this limits: the same scope as limitations[2] - what aspect of a clip drove the cell is undetermined Quotation withheld because this record is not cleared for redistribution. whether the several clips one neuron responds to share any abstract rule is left open duplicate of: limitations[4] · what this limits: the same scope as limitations[4] - the abstract-rule question is open Quotation withheld because this record is not cleared for redistribution. the inhibitory half of the effect could not be resolved in single cells because baseline firing is too low to see a decrease what this limits: the inhibitory half of the effect: BECAUSE OF THE LOW BASELINE FIRING RATE, THESE INHIBITORY RESPONSES WERE EVIDENT ONLY AT THE POPULATION LEVEL Quotation withheld because this record is not cleared for redistribution. the place-cell response profile — sparse firing from a very low baseline, sharply selective for one condition — imported as the expected shape of a human episode-selective response Quotation withheld because this record is not cleared for redistribution. the claim that navigational machinery was conserved and repurposed, used to license reading a spatial-memory literature as a memory literature Quotation withheld because this record is not cleared for redistribution. the analysis re-run with the questionable units removed, and the result reported unchanged eliminates: the reactivation effect resting on units whose classification is doubtful - the re-run without them is what shows it does not · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. response — sustained: elevated firing through most of the clip, up to 12 s Quotation withheld because this record is not cleared for redistribution. response — transient: elevation only during particular segments of a clip Quotation withheld because this record is not cleared for redistribution. response — the brief 300–600 ms MTL response of the prior stationary-stimulus literature, the sense this paper departs from Quotation withheld because this record is not cleared for redistribution. episode — the general cognitive object: multimodal representations linked in temporal succession Quotation withheld because this record is not cleared for redistribution. episode — the experimental stimulus, i.e. what a single 5-to-10-s clip depicts Quotation withheld because this record is not cleared for redistribution. episode — a sub-part contained within one clip, the grain a transient response may be picking out Quotation withheld because this record is not cleared for redistribution. unit — single unit (one isolated neuron) versus multi unit (unresolved cluster); both are counted together as units throughout Quotation withheld because this record is not cleared for redistribution. responsive — defined by consistency across all trials of a clip, not by response size Quotation withheld because this record is not cleared for redistribution. recall — the internal event, versus the verbal report that makes it observable; the paper keeps these apart and measures the lag between them 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: single-unit and multi-unit recording from clinically placed depth electrodes in epilepsy patients, across a viewing session and a later free-recall session 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: free recall — the spontaneous, internally generated emergence of an episodic memory into consciousness a second reading DIFFERING from the recorded value. Recorded: single-neuron reactivation during free recall — 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. extends → rodent studies of internally generated hippocampal replay during sleep, rest and waking (refs 27-33) Quotation withheld because this record is not cleared for redistribution. extends → prior reports of selective single-unit activation during mental imagery of stationary stimuli (ref 25) Quotation withheld because this record is not cleared for redistribution. extends → rodent work on internally generated hippocampal replay (refs 27–33), which could not be tied to recall because the animal cannot declare a memory differs in: THE AXIS IS WHETHER REPLAY HAS EVER BEEN TIED TO RECALL. 'Quotation withheld' The rodent replay literature is not disputed; the LINK to recall is called conjecture - and this record's [0] gives the reason it had to be, since free recall requires a declaration an animal cannot make. The limitation and its cause, in two rows of one paper. Quotation withheld because this record is not cleared for redistribution. extends → neuroimaging studies of reinstatement at recall (refs 21–24), whose resolution could not reach single neurons status: the paper's own words: indirectly measure Quotation withheld because this record is not cleared for redistribution. contradicts → the prior single-unit picture of MTL responses as brief, 300–600 ms events that rarely outlast 1 to 2 s Quotation withheld because this record is not cleared for redistribution. extends → prior single-unit imagery work (ref 25), which used an external content-specific cue Quotation withheld because this record is not cleared for redistribution. builds-on → the established finding that human MTL neurons respond selectively to complex stimuli, categories and individual persons or landmarks (refs 5–11) Quotation withheld because this record is not cleared for redistribution. free recall —contrasted-with→ recognition memory free recall recognition memory differs in: THE AXIS IS WHETHER THE EVENT IS OBSERVABLE TO THE EXPERIMENTER. 'Whereas the neuronal basis of recognition memory can be probed experimentally in human and nonhuman primates, Quotation withheld.' The two are not distinguished by any property of memory - they are distinguished by WHETHER THE EXPERIMENTER CAN SEE THE EVENT HAPPEN, so the division is one of experimental access and it explains why one literature is larger than the other. · strength of the relation: 0.95 Quotation withheld because this record is not cleared for redistribution. RETRACTED (NO RENAMING, AND THE SECOND ENDPOINT IS NOT IN THE SPAN) · replay —renamed-across-fields→ reactivation replay reactivation RETRACTED: NO RENAMING, AND THE SECOND ENDPOINT IS NOT IN THE SPAN. The edge is 'replay --renamed-across-fields--> reactivation'. The span reads: 'now DIRECTLY…' It names 'replay' and links it to FREE RECALL - not to 'reactivation', which does not appear at all. A renaming names BOTH names; this names one and relates it to a third thing. Found by reading the whole 'renamed-across-fields' untested population, 21 of 21, where it is one of only three genuinely weak edges against ~18 that were sound. — strength of the relation: 0.65 — the record marks this weak Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span links recall to neuronal replay; renamed-across-fields 'reactivation' is loose. Quotation withheld because this record is not cleared for redistribution. RETRACTED (SAME FAULT, SAME PAPER, DIFFERENT ENDPOINT) · reinstated —renamed-across-fields→ reactivation reinstated reactivation RETRACTED: SAME FAULT, SAME PAPER, DIFFERENT ENDPOINT. The edge is 'reinstated --renamed-across-fields--> reactivation'. The span - 'the BOLD signal…' - names 'reinstated' and describes when it happens. 'reactivation' is absent. Both this and [1] point at the same missing term, which suggests the tagger had a synonym set in mind that the paper never states in one place. WHAT THE SPAN DOES SAY: that learning-period activity is reinstated at recall, measured by BOLD, with the fMRI caveat attached - a phenomenon, not a terminological equivalence, and it is already carried by other tags on this record. — strength of the relation: 0.6 — the record marks this weak · note: endpoint was 'BOLD signal', which is a measurement modality, not another NAME for the phenomenon. The renaming the quote reports is reinstated/reactivation. Caught by querying the corpus for the cross-vocabulary link this edge was recorded to provide, and finding it did not bridge. Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span says learning activity is reinstated at recall; renamed-across-fields is loose. Quotation withheld because this record is not cleared for redistribution. sustained response —distinct-from→ transient sustained response transient strength of the relation: 0.9 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span writes 'in a sustained or transient manner'; it lists the two, it does not distinguish them. Quotation withheld because this record is not cleared for redistribution. reactivation —operationalised-as→ FR ratio reactivation FR ratio strength of the relation: 0.85 Audit 2026-09-28: DOUBTFUL - read against its span and kept, but reactivation was measured as a firing-rate ratio; operationalised-as is loose. Quotation withheld because this record is not cleared for redistribution. reactivation —is-a→ neuronal correlate reactivation neuronal correlate strength of the relation: 0.75 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the authors 'SUGGEST' the reactivation is a neuronal correlate of recollection; is-a is hedged. Quotation withheld because this record is not cleared for redistribution. mental imagery —contrasted-with→ free recall mental imagery free recall strength of the relation: 0.85 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but endpoint by anaphora: 'our study' = free recall (previous sentences); contrasted-with is loose. Quotation withheld because this record is not cleared for redistribution. hippocampus —part-of→ medial temporal lobe hippocampus medial temporal lobe strength of the relation: 0.85 Quotation withheld because this record is not cleared for redistribution. EVIDENCE CORRECTED (RE-GROUNDED) · verbal report —operationalised-as→ free recall verbal report free recall EVIDENCE CORRECTED: RE-GROUNDED. The old span named neither endpoint. This sentence names both and states how the measure was taken - a verbal report AT THE TIME OF free recall. — strength of the relation: 0.9 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but verbal reports of memories were given AT THE TIME OF free recall; operationalised-as is loose. Quotation withheld because this record is not cleared for redistribution. chi-squared chi-squared 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. standard error of the mean standard error of the mean Quotation withheld because this record is not cleared for redistribution. standard deviation (used as a response threshold) standard deviation (used as a response threshold) Quotation withheld because this record is not cleared for redistribution. gaussian kernel smoothing gaussian kernel smoothing Quotation withheld because this record is not cleared for redistribution. two-sample t test two-sample t test Quotation withheld because this record is not cleared for redistribution. Student’s t test Student’s t test Quotation withheld because this record is not cleared for redistribution. χ 2 χ 2 Quotation withheld because this record is not cleared for redistribution. χ 2 χ 2 Quotation withheld because this record is not cleared for redistribution. SEM SEM Quotation withheld because this record is not cleared for redistribution. SD SD Quotation withheld because this record is not cleared for redistribution. post–stimulus time histogram post–stimulus time histogram Quotation withheld because this record is not cleared for redistribution. A subset of human medial temporal lobe neurons fired selectively and in a sustained way throughout and after specific video episodes, for durations up to 12 seconds. Quotation withheld because this record is not cleared for redistribution. The same hippocampal and entorhinal neurons reactivated before the patient verbally reported freely recalling the corresponding episode. Quotation withheld because this record is not cleared for redistribution. Reactivation at recall appeared in the hippocampal and entorhinal populations but not in medial frontal units. Quotation withheld because this record is not cleared for redistribution. In an individual recall trial, firing rose above baseline roughly 1.5 seconds before the verbal report and peaked shortly before speech onset, preceding rather than following the report. Quotation withheld because this record is not cleared for redistribution. Episode-specific reactivation was weak in amygdala and absent from parahippocampal gyrus. Quotation withheld because this record is not cleared for redistribution. 54.9% of recorded units (475) responded significantly to at least one clip. Quotation withheld because this record is not cleared for redistribution. Only 27% of units were recorded within the epileptogenic seizure foci. Quotation withheld because this record is not cleared for redistribution. 475 units (54.9%) responded to at least one clip Quotation withheld because this record is not cleared for redistribution. sustained responses were almost entirely an MTL phenomenon: 44 of the 46 such cells were in MTL, only 2 in medial frontal lobe Quotation withheld because this record is not cleared for redistribution. responses lasted as long as 12 s and were typically ended only by the onset of the next clip Quotation withheld because this record is not cleared for redistribution. across the population, firing rose above baseline in the 3 s before the verbal report and stayed elevated for 2 s after it Quotation withheld because this record is not cleared for redistribution. the effect was present in hippocampal and entorhinal units and absent in medial frontal units Quotation withheld because this record is not cleared for redistribution. reactivation was strongest in cells whose viewing response had been sustained rather than transient Quotation withheld because this record is not cleared for redistribution. within MTL the effect was weak in amygdala, absent in parahippocampal gyrus, and striking for hippocampal and entorhinal neurons with sustained responses Quotation withheld because this record is not cleared for redistribution. the proportion of responsive units did not differ across the sampled regions Quotation withheld because this record is not cleared for redistribution. removing the 27% of units recorded inside the seizure focus did not change the result Quotation withheld because this record is not cleared for redistribution. cells stayed at baseline during recall of clips that had not driven them at viewing Quotation withheld because this record is not cleared for redistribution. the population firing was already slightly elevated before clip onset, which the authors read as anticipation Quotation withheld because this record is not cleared for redistribution. patients recalled a mean of 83.2% of the clips they had been shown Quotation withheld because this record is not cleared for redistribution. responsive unit responsive unit Quotation withheld because this record is not cleared for redistribution. sustained response sustained response Quotation withheld because this record is not cleared for redistribution. episode episode Quotation withheld because this record is not cleared for redistribution. responsive responsive Quotation withheld because this record is not cleared for redistribution. sustained response sustained response Quotation withheld because this record is not cleared for redistribution. transient transient Quotation withheld because this record is not cleared for redistribution. FR ratio FR ratio Quotation withheld because this record is not cleared for redistribution. episodes episodes Quotation withheld because this record is not cleared for redistribution. baseline firing rate baseline firing rate Quotation withheld because this record is not cleared for redistribution. place cells place cells Quotation withheld because this record is not cleared for redistribution. instantaneous firing rate instantaneous firing rate Quotation withheld because this record is not cleared for redistribution. clips with nonsignificant responses clips with nonsignificant responses Quotation withheld because this record is not cleared for redistribution. firing rate firing rate Quotation withheld because this record is not cleared for redistribution. baseline firing rate baseline firing rate Quotation withheld because this record is not cleared for redistribution. FR ratio FR ratio Quotation withheld because this record is not cleared for redistribution. onset time of the verbal report of recall onset time of the verbal report of recall Quotation withheld because this record is not cleared for redistribution. preferred clip preferred clip Quotation withheld because this record is not cleared for redistribution. clip onset and offset clip onset and offset Quotation withheld because this record is not cleared for redistribution. sound amplitude of patient voice sound amplitude of patient voice Quotation withheld because this record is not cleared for redistribution. 857 units (441 single units and 416 multi units) — Total units recorded across patients in medial temporal lobe and medial frontal cortex Quotation withheld because this record is not cleared for redistribution. 13 patients, 43 sessions — Sample size of epilepsy patients and of paired viewing and recall sessions Quotation withheld because this record is not cleared for redistribution. 83.2% (±5% SEM) — Mean proportion of presented clips that patients spontaneously recalled Quotation withheld because this record is not cleared for redistribution. 475 units (54.9%) — Units showing a significant response to one or more clips during viewing Quotation withheld because this record is not cleared for redistribution. 46 (9.7%) — Responsive units showing a sustained response to at least one clip Quotation withheld because this record is not cleared for redistribution. 44 versus 2, P < 0.03 — Distribution of the 46 sustained-response cells between medial temporal lobe and medial frontal lobe Quotation withheld because this record is not cleared for redistribution. 15.57 Hz versus 2.11 and 2.23 Hz — Firing rate of the entorhinal unit during its preferred Simpsons clip versus other clips and blank periods Quotation withheld because this record is not cleared for redistribution. 17.7% ± 0.7 SEM (or 4.0 ± 0.2 SEM clips) — Average selectivity of responsive units: share of presented clips eliciting a sustained or transient response Quotation withheld because this record is not cleared for redistribution. 1.4 ± 0.1 SEM clips (5.7% ± 0.5 SEM) — Average number of clips to which a sustained-response unit responded in a sustained manner Quotation withheld because this record is not cleared for redistribution. 27% — Share of recorded units that lay within the epileptogenic seizure foci Quotation withheld because this record is not cleared for redistribution. 857 units — total units recorded, 441 single and 416 multi Quotation withheld because this record is not cleared for redistribution. 475 units (54.9%) — units responding to one or more clips Quotation withheld because this record is not cleared for redistribution. 46 (9.7%) — responsive units showing a sustained response Quotation withheld because this record is not cleared for redistribution. 44 of 46 — sustained-response cells located in MTL rather than medial frontal lobe Quotation withheld because this record is not cleared for redistribution. 12 s — longest observed response duration Quotation withheld because this record is not cleared for redistribution. 15.57 Hz — entorhinal unit firing rate to the Simpsons clip, against 2.11 and 2.23 Hz for other clips and blanks Quotation withheld because this record is not cleared for redistribution. 1500 ms — lead of the firing-rate rise over the onset of verbal report Quotation withheld because this record is not cleared for redistribution. 100 ms — lead of the firing peak over verbal report onset Quotation withheld because this record is not cleared for redistribution. 13 patients, 43 sessions — the sample Quotation withheld because this record is not cleared for redistribution. 83.2% (±5% SEM) — proportion of presented clips spontaneously recalled Quotation withheld because this record is not cleared for redistribution. 17.7% ± 0.7 SEM — share of presented clips a responsive unit responded to, sustained or transient Quotation withheld because this record is not cleared for redistribution. 1.4 ± 0.1 SEM clips — clips per unit driving a sustained response, i.e. 5.7% of those presented Quotation withheld because this record is not cleared for redistribution. 27% — share of units recorded inside the epileptogenic seizure foci Quotation withheld because this record is not cleared for redistribution. 300 and 600 ms — typical duration of MTL responses in the prior literature Quotation withheld because this record is not cleared for redistribution. 5 to 10 s — duration of each audiovisual clip Quotation withheld because this record is not cleared for redistribution. n = 154 — responsive hippocampal and entorhinal cells in the population average Quotation withheld because this record is not cleared for redistribution. n = 56 — anterior cingulate cells in the control population average Quotation withheld because this record is not cleared for redistribution. 20 cells — sustained cells holding elevated firing at least 1 s past clip offset Quotation withheld because this record is not cleared for redistribution. 48 clips — clip set shown to the patient of the entorhinal example Quotation withheld because this record is not cleared for redistribution. 3 SD — threshold above session mean firing used to call a recall-time elevation Quotation withheld because this record is not cleared for redistribution. 1200 ms — full width at half maximum of the smoothing Gaussian Quotation withheld because this record is not cleared for redistribution. 1 to 5 min — duration of the intervening task between viewing and recall Quotation withheld because this record is not cleared for redistribution. 500-ms — bin width for histograms and the minimum window for calling a response 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.1164685.