Percepts to recollections: insights from single neuron recordings in the human brain

Percepts to recollections: insights from single neuron recordings in the human brain Suthana 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.

Percepts to recollections: insights from single neuron recordings in the human brain 2012 Trends in Cognitive Sciences 10.1016/j.tics.2012.06.006 Suthana2012-percepts-to-recollections-insights-from-sing Suthana2012-percepts-to-recollections-insights-from-sing.txt, the project's staged copy b70bb5ab6a6b59ddfc44ae5e24269c6e764b93df 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 response properties of single medial temporal lobe neurons in humans, from perception to recollection Quotation withheld because this record is not cleared for redistribution. what neuronal code carries the transformation from sensory information to consciously retrievable episodic memory Quotation withheld because this record is not cleared for redistribution. collects the properties of human MTL neurons - selective, invariant, late, tied to conscious perception, internally generatable, reactivated before recall - into one account of an explicit, abstracted code Quotation withheld because this record is not cleared for redistribution. existence Quotation withheld because this record is not cleared for redistribution. single-neuron coding in the human medial temporal lobe during perception and episodic memory Quotation withheld because this record is not cleared for redistribution. review review Quotation withheld because this record is not cleared for redistribution. mechanisms MTL neurons transform novel percepts into sparse, abstracted-yet-detailed representations, and it is this abstraction that later allows the representation to be summoned by either internal or external cues, producing conscious episodic recall Quotation withheld because this record is not cleared for redistribution. mechanisms Representations must carry both detail and abstraction (loss of detail) so that a partial internal or external cue is sufficient to reactivate them Quotation withheld because this record is not cleared for redistribution. mechanisms Entorhinal stimulation enhances memory by resetting the phase of the hippocampal theta rhythm, aligning encoding with a favourable oscillatory phase Quotation withheld because this record is not cleared for redistribution. the hippocampus sits at the top of a multisensory hierarchy, receiving object information via perirhinal and spatial information via parahippocampal cortex Quotation withheld because this record is not cleared for redistribution. invariance crosses sensory modality: the same neuron responds to a concept presented as image, sound or written word Quotation withheld because this record is not cleared for redistribution. selective responses appear only when the subject consciously recognises the stimulus, and are near-silent when recognition fails Quotation withheld because this record is not cleared for redistribution. imagining a stimulus recruits the same neuron that fires when it is seen, and subjects can steer a display by modulating two such neurons Quotation withheld because this record is not cleared for redistribution. hippocampal and entorhinal neurons reactivate just before a spontaneous verbal report of recall, and only those regions do Quotation withheld because this record is not cleared for redistribution. firing-rate correlation across successive time points within an episode grows with repetition and predicts later recall, proposed as temporal binding Quotation withheld because this record is not cleared for redistribution. entorhinal white-matter stimulation enhances memory and resets hippocampal theta phase, offered as the mechanism Quotation withheld because this record is not cleared for redistribution. The human single-neuron dataset is rare and constrained by what clinical circumstances permit. what this limits: the evidence base for human single-neuron work: THE DATA PRESENTED ARE RARE AND LIMITED BY CLINICAL OPPORTUNITIES Quotation withheld because this record is not cleared for redistribution. Recordings are only possible where electrodes are implanted for clinical reasons, restricting the tasks and settings that can be studied. what this limits: any generalisation beyond the implanted population: STUDIES ARE DONE ONLY IN SPECIAL CIRCUMSTANCES WHERE ELECTRODES ARE PLACED IN THE BRAIN FOR SPECIFIC CLINICAL REASONS Quotation withheld because this record is not cleared for redistribution. Sampling is biased to anterior and middle hippocampus, so anterior-posterior comparisons cannot yet be made. what this limits: anterior-posterior comparisons along the hippocampal axis: SYSTEMATIC ANTERIOR-POSTERIOR ANALYSIS IN THE HIPPOCAMPAL AXIS WILL HAVE TO AWAIT FUTURE STUDIES Quotation withheld because this record is not cleared for redistribution. No human perirhinal neurons have been recorded, so the proposed hippocampus/perirhinal recollection-familiarity dissociation cannot be tested. what this limits: the proposed hippocampal/perirhinal dissociation: HUMAN SINGLE NEURON STUDIES TO DATE HAVE NOT RECORDED FROM PERIRHINAL NEURONS, so FUTURE STUDIES ARE NEEDED TO INVESTIGATE THIS DISSOCIATION Quotation withheld because this record is not cleared for redistribution. correlation of single-neuron responses with conscious percepts does not establish that the region is necessary for them what this limits: necessity claims specifically: THE CORRELATION OF SINGLE NEURON RESPONSES IN THE MTL WITH SPECIFIC CONSCIOUS PERCEPTS DOES NOT IMPLY THAT THESE REGIONS ARE NECESSARY FOR CONSCIOUS PERCEPTS Quotation withheld because this record is not cleared for redistribution. most stimuli used were highly familiar to the participant, so the responses may partly reflect familiarity what this limits: the studies that used familiar stimuli, and the familiarity reading of their responses: THEY OFTEN USED STIMULI THAT WERE HIGHLY FAMILIAR TO THE PARTICIPANT (E.G. A FAMILY MEMBER OR FAMOUS PERSON/LANDMARK) Quotation withheld because this record is not cleared for redistribution. the recognition-related responses were not stimulus-specific and so need not represent single memories what this limits: single-memory claims: THE RESPONSES IN THE PREVIOUSLY DISCUSSED STUDIES WERE NOT SPECIFIC TO INDIVIDUAL STIMULI AND THUS DO NO NECESSARILY REPRESENT SINGLE MEMORIES Quotation withheld because this record is not cleared for redistribution. no human single-neuron study has recorded from perirhinal cortex, leaving the key dissociation untested what this limits: the same scope as limitations[3] - the hippocampal/perirhinal recollection-familiarity dissociation, untested because no human perirhinal neurons have been recorded Quotation withheld because this record is not cleared for redistribution. recordings come only from patients confined to a hospital room who cannot navigate a real environment what this limits: real-world navigation studies: PATIENTS LIMITED TO THEIR HOSPITAL ROOM WHO CANNOT NAVIGATE A REAL ENVIRONMENT DURING RECORDINGS Quotation withheld because this record is not cleared for redistribution. fewer than 100 neurons are recorded at once, scattered across regions chosen on clinical grounds what this limits: any population-level claim: Quotation withheld AND THESE NEURONS MAY BE DISTRIBUTED ACROSS SEVERAL BRAIN AREAS, WHICH NEED TO BE CLINICALLY SAMPLED Quotation withheld because this record is not cleared for redistribution. spikes on one microelectrode probably come from several neighbouring neurons, especially in densely packed hippocampus what this limits: single-neuron claims in dense structures: SPIKES DETECTED ON A SINGLE MICROELECTRODE MOST LIKELY ORIGINATE FROM SEVERAL NEIGHBORING NEURONS, ESPECIALLY WITHIN BRAIN AREAS THAT HAVE DENSELY PACKED CELLULAR STRUCTURES, SUCH AS THE HIPPOCAMPUS Quotation withheld because this record is not cleared for redistribution. recordings are almost all anterior and middle hippocampus, so the anterior-posterior axis cannot yet be analysed duplicate of: limitations[2] · what this limits: the same scope as limitations[2] - the anterior-posterior axis, unanalysable while recordings are ANTERIOR AND MIDDLE only Quotation withheld because this record is not cleared for redistribution. electrode bundles can be localised to about 1 mm but individual electrodes within a bundle cannot be separated what this limits: any localisation finer than the bundle: BUNDLES OF MICROELECTRODES CAN BE LOCALIZED TO AREAS AS SMALL AS 1mm, ALTHOUGH SEPARATION OF A SINGLE MICROELECTRODE FROM ITS BUNDLED NEIGHBORING ELECTRODES IS STILL A CHALLENGE Quotation withheld because this record is not cleared for redistribution. the place cell, transferred twice - literally to human virtual navigation, and metaphorically to concept coding in a cognitive multidimensional space Quotation withheld because this record is not cleared for redistribution. reactivation of a fear memory by stimulating specific hippocampal neurons, held up as the model for a causal human approach Quotation withheld because this record is not cleared for redistribution. invariant here means invariant across physical form AND across sensory modality - image, sound, written name - not merely across viewpoint Quotation withheld because this record is not cleared for redistribution. place cell is used twice over: literally for rodent and human spatial cells, and metaphorically for concept cells as place cells in a cognitive multidimensional space Quotation withheld because this record is not cleared for redistribution. recollection versus familiarity is a behavioural distinction operationalised by whether spatial position is also recalled; the hippocampal data are read as tracking memory strength instead Quotation withheld because this record is not cleared for redistribution. declarative memory is defined here by what the SUBJECT can report, which is why the paper argues the question belongs to humans Quotation withheld because this record is not cleared for redistribution. ultra sparse coding, asserted for MTL - note that Rigotti2013 reports PFC representations are rather dense, a different region and the opposite value of the same construct 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: narrative review of intracranial single-neuron recordings in epilepsy patients implanted for clinical reasons a second reading DIFFERING from the recorded value. Recorded: review — 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: concept-level coding in the human medial temporal lobe a second reading DIFFERING from the recorded value. Recorded: single-neuron coding in the human medial temporal lobe during perception and episodic memory — 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. cites → Gelbard-Sagiv et al. on hippocampal reactivation before free recall - a paper independently tagged in this corpus Quotation withheld because this record is not cleared for redistribution. supports → the view that MTL neurons code concepts explicitly and invariantly Quotation withheld because this record is not cleared for redistribution. qualifies → the inference from selective responses to necessity for conscious perception Quotation withheld because this record is not cleared for redistribution. supports → the account of hippocampus as tracking memory strength rather than distinguishing recollection from familiarity Quotation withheld because this record is not cleared for redistribution. extends → rodent place-cell physiology, by proposing concept cells as place cells in a cognitive space Quotation withheld because this record is not cleared for redistribution. EVIDENCE CORRECTED (RE-GROUNDED) · recollection —contrasted-with→ familiarity recollection familiarity EVIDENCE CORRECTED: RE-GROUNDED. The old span named neither. This clause names both and sets them as the two mechanisms a recognition task can run on. — strength of the relation: 1 · note: and the hippocampal data are read as NOT respecting the distinction Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span says the hippocampus does not differentiate recollection from familiarity; contrasted-with is loose. Quotation withheld because this record is not cleared for redistribution. EVIDENCE CORRECTED (RE-GROUNDED) · theta —causes→ episodic memory theta episodic memory EVIDENCE CORRECTED: RE-GROUNDED. The old span described the coordination without naming episodic memory. This sentence names both and states the claim the edge records - 'implicate A DIRECT ROLE FOR theta-linked spiking activity…' - which is a stronger causal reading than the old span supported. — status: PREDICTIVE, NOT DEMONSTRATED CAUSAL. The span reports that the tighter spike-theta coordination PREDICTED which memories succeeded; nothing was manipulated, so the direction of the arrow is an inference the paper does not test. · strength of the relation: 0.7 · note: predictive rather than demonstrated causal; the weight carries that Audit 2026-09-28: DOUBTFUL - read against its span and kept, but theta-linked spiking predicted subsequent memory; causes overstates a prediction. Quotation withheld because this record is not cleared for redistribution. recollection —contrasted-with→ familiarity recollection familiarity strength of the relation: 1 · note: RE-GROUNDS another row of this record: the relation was always stated, the endpoint names were not in the span. Quotation withheld because this record is not cleared for redistribution. theta —causes→ episodic memory theta episodic memory strength of the relation: 0.7 · note: RE-GROUNDS another row of this record: the old span named ONE endpoint. Quotation withheld because this record is not cleared for redistribution. The selective responses of human MTL neurons are invariant not only across different pictures of the same concept but across sensory modality, firing to the spoken name, written name or image of the same entity. Quotation withheld because this record is not cleared for redistribution. Selective and invariant MTL responses have relatively long latencies, often 300-500 ms. Quotation withheld because this record is not cleared for redistribution. MTL neurons selective for an image significantly increase firing only if the subject recognises it, even at 33 ms exposure. Quotation withheld because this record is not cleared for redistribution. Subjects can volitionally drive the firing of individual MTL neurons to determine which of two competing images is displayed, overriding the actual visual input on the screen. Quotation withheld because this record is not cleared for redistribution. Hippocampal and entorhinal neurons selectively active during the initial viewing of a film episode were reactivated just before the subject verbally reported recalling it. Quotation withheld because this record is not cleared for redistribution. Tighter phase coordination between hippocampal spiking and the theta LFP during encoding predicts whether that particular image will later be remembered. Quotation withheld because this record is not cleared for redistribution. Growing correlation of a hippocampal neuron's firing across successive time points within an episode, built up over repetitions, predicts later recall of that episode. Quotation withheld because this record is not cleared for redistribution. Single hippocampal neurons distinguished new from old stimuli after a single trial, independently of whether spatial position was recalled. Quotation withheld because this record is not cleared for redistribution. Deep-brain stimulation of the entorhinal white matter during learning enhances later recall, whereas stimulation of the hippocampus itself does not. Quotation withheld because this record is not cleared for redistribution. MTL neurons respond to a concept across radically different images and even its written name Quotation withheld because this record is not cleared for redistribution. the invariance holds across sensory modality - image, audio, written word Quotation withheld because this record is not cleared for redistribution. selective responses are late, 300-500 ms, and longer than those seen in animal studies Quotation withheld because this record is not cleared for redistribution. a neuron fires for its concept only when the subject reports recognising it, even at 33 ms exposure Quotation withheld because this record is not cleared for redistribution. imagining a stimulus recruits the same neuron as seeing it Quotation withheld because this record is not cleared for redistribution. subjects steered a display by modulating two independently selective neurons, overriding the visual input Quotation withheld because this record is not cleared for redistribution. hippocampal and entorhinal neurons reactivate before a spontaneous report of recall; amygdala and frontal cortex do not Quotation withheld because this record is not cleared for redistribution. the reactivation was visible in a single free-recall trial in a single neuron Quotation withheld because this record is not cleared for redistribution. new selective responses to experimenters formed over hours or days Quotation withheld because this record is not cleared for redistribution. several rodent cell types - grid, border, head-direction, time - have not yet been characterised in humans Quotation withheld because this record is not cleared for redistribution. Episodic memory Episodic memory Quotation withheld because this record is not cleared for redistribution. Medial temporal lobe (MTL) Medial temporal lobe (MTL) Quotation withheld because this record is not cleared for redistribution. Spikes Spikes Quotation withheld because this record is not cleared for redistribution. Local field potentials (LFPs) Local field potentials (LFPs) Quotation withheld because this record is not cleared for redistribution. Recollection vs. familiarity Recollection vs. familiarity Quotation withheld because this record is not cleared for redistribution. invariant response invariant response Quotation withheld because this record is not cleared for redistribution. declarative memory declarative memory Quotation withheld because this record is not cleared for redistribution. spike sorting spike sorting Quotation withheld because this record is not cleared for redistribution. display time display time Quotation withheld because this record is not cleared for redistribution. firing rate firing rate Quotation withheld because this record is not cleared for redistribution. spike-field coherence spike-field coherence Quotation withheld because this record is not cleared for redistribution. subsequently remembered versus forgotten subsequently remembered versus forgotten Quotation withheld because this record is not cleared for redistribution. 300 msec - 500 msec — typical latency range of selective, invariant MTL single-neuron responses Quotation withheld because this record is not cleared for redistribution. 33 to 256 msec — range of image display durations used to dissociate conscious recognition from stimulus presence Quotation withheld because this record is not cleared for redistribution. 3–8Hz — frequency band of the theta LFP oscillation implicated in human memory Quotation withheld because this record is not cleared for redistribution. 30–100 Hz — frequency band of the gamma oscillation linked to the recall-predicting theta amplitude Quotation withheld because this record is not cleared for redistribution. 100 images (50 old and 50 new stimuli) — size and composition of the recognition-memory test set in the spike-theta encoding study Quotation withheld because this record is not cleared for redistribution. 7 out of 8 — trials on which one example subject successfully faded the target actor's image by controlling neuronal firing Quotation withheld because this record is not cleared for redistribution. less than 100 — number of neurons that can currently be recorded simultaneously in a human depth-electrode setup Quotation withheld because this record is not cleared for redistribution. 300 to 500 ms — latency of selective invariant responses Quotation withheld because this record is not cleared for redistribution. 33 to 256 ms — the display-time range over which recognition gates the response Quotation withheld because this record is not cleared for redistribution. 3 to 8 Hz theta — the oscillation predicting encoding success Quotation withheld because this record is not cleared for redistribution. fewer than 100 neurons simultaneously — the ceiling of the human recording setup Quotation withheld because this record is not cleared for redistribution. about 1 mm — the precision to which an electrode bundle can be localised Quotation withheld because this record is not cleared for redistribution. spike-field coherence spike-field coherence 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/10.1016/j.tics.2012.06.006.