Spatially Distributed Local Fields in the Hippocampus Encode Rat Position

Spatially Distributed Local Fields in the Hippocampus Encode Rat Position Agarwal Stevenson Berényi Mizuseki Buzsáki Sommer

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.

Spatially Distributed Local Fields in the Hippocampus Encode Rat Position 2014 Science 10.1126/science.1250444 Agarwal2014-spatially-distributed-local-fields-in-the-hi Agarwal2014-spatially-distributed-local-fields-in-the-hi.txt, the project's staged copy c57f4fd82e1575e7b15a34638010b4dd1c0de8fa 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 hippocampal local field potential in freely moving rats, and whether the animal's position can be read out of it. Quotation withheld because this record is not cleared for redistribution. If no single electrode's field potential is tuned to place, can position nevertheless be recovered from many electrodes at once? Quotation withheld because this record is not cleared for redistribution. Shows the spatiotemporal structure of the theta rhythm encodes position about as well as spiking populations do, and introduces feature-tuned field potentials as the LFP counterpart of place cells. Quotation withheld because this record is not cleared for redistribution. existence Quotation withheld because this record is not cleared for redistribution. spatial position coding in hippocampal local field potentials Quotation withheld because this record is not cleared for redistribution. electrophysiology electrophysiology Quotation withheld because this record is not cleared for redistribution. mechanisms Because hippocampal place cells are not topographically arranged, any one electrode pools dissimilarly tuned neurons and therefore shows only weak place modulation, which is why the code appears only across many sites Quotation withheld because this record is not cleared for redistribution. mechanisms During movement the population is driven by a single sparse cause - current position - so unsupervised source separation can recover that cause from mixed, subsampled field-potential measurements Quotation withheld because this record is not cleared for redistribution. mechanisms The theta LFP inherits position information because it is generated mainly by the membrane potential fluctuations of the place cell population itself Quotation withheld because this record is not cleared for redistribution. mechanisms the theta rhythm is treated as a radio carrier wave, and position is recovered by demodulating it - stripping the carrier phase to leave a spatial pattern Quotation withheld because this record is not cleared for redistribution. mechanisms the LFP carries place information because it is generated by the membrane potentials of the place-cell population itself duplicate of: mechanisms[2] Quotation withheld because this record is not cleared for redistribution. mechanisms position is a single sparse cause driving the whole population, which is why an unsupervised method can recover it without being told where the rat is Quotation withheld because this record is not cleared for redistribution. The unsupervised (ICA) discovery of position-tuned field-potential features degrades as trial-by-trial neuronal variability grows: fewer components are recovered from noisy populations. what this limits: the number of position-tuned components recovered: AS THE TRIAL-BY-TRIAL VARIABILITY OF NEURONAL ACTIVITY WAS INCREASED, ICA IDENTIFIED A DECREASING NUMBER OF FFPs Quotation withheld because this record is not cleared for redistribution. Each electrode pools a large, non-topographically-ordered population, so any single site is only weakly place-modulated. what this limits: any single-site interpretation of the field potential: EACH ELECTRODE SUBSAMPLES (POOLS) THE ACTIVITY OF A LARGE POPULATION OF NEURONS Quotation withheld because this record is not cleared for redistribution. The method degrades with neuronal variability: noisier populations yield fewer recoverable components. duplicate of: limitations[0] · what this limits: the same scope as limitations[0] - the component count is partly a measure of the noise Quotation withheld because this record is not cleared for redistribution. Recovered component width is bounded by the tuning width of the underlying population, so the readout cannot be sharper than what generates it. what this limits: the spatial resolution of the readout: FOR NOISY POPULATIONS, FFP WIDTH INCREASED IN PROPORTION TO, BUT NEVER EXCEEDED, THE TUNING WIDTH OF THE UNDERLYING POPULATION'S PLACE FIELDS Quotation withheld because this record is not cleared for redistribution. The mechanistic account of how FFPs relate to neuronal populations rests on a simulation, not on measurement. what this limits: the mechanistic account linking components to populations: A SIMULATION WAS CONSTRUCTED IN WHICH THE LFP WAS MODELED AS THE SUPERIMPOSED ACTIVITY GENERATED BY NUMEROUS PLACE-MODULATED SYNAPTIC INPUTS Quotation withheld because this record is not cleared for redistribution. The carrier-wave and demodulation model: theta is treated as a carrier whose modulation carries the message, and the analysis strips the carrier phase to read that message out Quotation withheld because this record is not cleared for redistribution. The compressed-sensing result that sparse signals can be recovered from mixed, subsampled measurements, used to justify recovering position from pooled field potentials by unsupervised learning Quotation withheld because this record is not cleared for redistribution. the carrier-and-modulation model of a signal, and the demodulation operation that recovers the message from it, applied to a brain rhythm Quotation withheld because this record is not cleared for redistribution. the result that a sparse signal is recoverable from mixed and subsampled measurements, used to justify reading position out of pooled field potentials duplicate of: transfers[1] Quotation withheld because this record is not cleared for redistribution. the LFP decoder benchmarked against established spike-based decoding on the same data, both cross-validated eliminates: a decoding accuracy with no reference point - the spike decoder is what the field already trusts, and the comparison is made on the same recordings with the same validation · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. local field potential local field potential Quotation withheld because this record is not cleared for redistribution. spikes spikes Quotation withheld because this record is not cleared for redistribution. theta rhythm theta rhythm Quotation withheld because this record is not cleared for redistribution. carrier wave carrier wave Quotation withheld because this record is not cleared for redistribution. demodulation demodulation Quotation withheld because this record is not cleared for redistribution. place cells place cells Quotation withheld because this record is not cleared for redistribution. feature-tuned field potentials feature-tuned field potentials Quotation withheld because this record is not cleared for redistribution. compressed sensing compressed sensing Quotation withheld because this record is not cleared for redistribution. independent component analysis independent component analysis Quotation withheld because this record is not cleared for redistribution. phase precession phase precession Quotation withheld because this record is not cleared for redistribution. carrier - in radio communication, the steady wave whose modulation carries the message; used here for the hippocampal theta oscillation Quotation withheld because this record is not cleared for redistribution. code - a distributed population code, where information is in the pattern across many sites rather than in any one of them Quotation withheld because this record is not cleared for redistribution. sparse - in the compressed-sensing sense: few active causes, which is what makes unsupervised recovery possible Quotation withheld because this record is not cleared for redistribution. phenomenon: distributed encoding of spatial position in the hippocampal local field potential a second reading DIFFERING from the recorded value. Recorded: spatial position coding in hippocampal local field potentials — 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 → earlier large-scale LFP recording work, cited only as reference 25 Quotation withheld because this record is not cleared for redistribution. builds-on → the compressed-sensing theoretical result cited as reference 14 (with compressed sensing, reference 15) Quotation withheld because this record is not cleared for redistribution. contradicts → the assumption that pooling across a non-topographic population destroys the information in an LFP Quotation withheld because this record is not cleared for redistribution. builds-on → the theta-phase precession literature on place cell spikes ( 19 ) Quotation withheld because this record is not cleared for redistribution. theta rhythm —renamed-across-fields→ carrier wave theta rhythm carrier wave strength of the relation: 1 Quotation withheld because this record is not cleared for redistribution. local field potential —distinct-from→ spikes local field potential spikes strength of the relation: 1 Quotation withheld because this record is not cleared for redistribution. place cells —causes→ local field potential place cells local field potential strength of the relation: 0.9 Quotation withheld because this record is not cleared for redistribution. feature-tuned field potentials —contrasted-with→ place cells feature-tuned field potentials place cells differs in: THE AXIS IS HOW MANY FIELDS EACH UNIT HAS. 'Quotation withheld.' The consequence is stated too: the field count is what makes the two populations separable, so the axis is also the reason the contrast matters. · strength of the relation: 1 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span writes 'FFPs' and 'pyramidal cells ... place fields', not 'place cells'. Quotation withheld because this record is not cleared for redistribution. theta rhythm —part-of→ local field potential theta rhythm local field potential strength of the relation: 1 Quotation withheld because this record is not cleared for redistribution. principal component analysis principal component analysis Quotation withheld because this record is not cleared for redistribution. independent component analysis independent component analysis Quotation withheld because this record is not cleared for redistribution. cross-validation cross-validation Quotation withheld because this record is not cleared for redistribution. median median Quotation withheld because this record is not cleared for redistribution. linear estimator (linear decoding) linear estimator (linear decoding) Quotation withheld because this record is not cleared for redistribution. Bayesian estimation Bayesian estimation Quotation withheld because this record is not cleared for redistribution. sparse coding (unsupervised dictionary learning) sparse coding (unsupervised dictionary learning) Quotation withheld because this record is not cleared for redistribution. median median Quotation withheld because this record is not cleared for redistribution. principal-component-analysis principal-component-analysis Quotation withheld because this record is not cleared for redistribution. variance variance Quotation withheld because this record is not cleared for redistribution. The spatiotemporal structure of the hippocampal theta rhythm, sampled across many electrodes, encodes the rat's position about as well as populations of spiking neurons do. Quotation withheld because this record is not cleared for redistribution. Decoding accuracy from field potentials matched that from spikes, with a slightly smaller median error, though the two decoders depended differently on running velocity and position. Quotation withheld because this record is not cleared for redistribution. Successful decoding required both removing the theta carrier phase and keeping the signal's high dimensionality, even though most of its variance lay in a low-dimensional subspace. Quotation withheld because this record is not cleared for redistribution. Unsupervised decomposition of the theta-band field potential recovered components that fire selectively at particular track positions and carry most of the signal variance. Quotation withheld because this record is not cleared for redistribution. Each FFP's phase advanced progressively as the animal crossed its field, mirroring theta phase precession of place cell spikes. Quotation withheld because this record is not cleared for redistribution. FFPs covered the track more evenly than pyramidal cell place fields did, which matched the relative accuracy of the two decoders. Quotation withheld because this record is not cleared for redistribution. Position could still be decoded and FFPs identified from dendritic-layer sites, excluding the somatic layers where spikes were detected. Quotation withheld because this record is not cleared for redistribution. A different, broadband sparse-coding analysis recovered position-tuned components spread along the whole track, matching the theta-band result. Quotation withheld because this record is not cleared for redistribution. In simulation, for noisy populations FFP width grew in proportion to, but never exceeded, the tuning width of the underlying place fields. Quotation withheld because this record is not cleared for redistribution. Position decoded from the LFP is about as accurate as position decoded from spikes - median error 6.7 cm against 9.2 cm. Quotation withheld because this record is not cleared for redistribution. A subset of the decomposed components, termed FFPs, is selectively active at specific track positions along one running direction. Quotation withheld because this record is not cleared for redistribution. These components carry most of the signal - 53 to 79% of total variance across three rats. Quotation withheld because this record is not cleared for redistribution. The ICA-derived FFPs show several place-cell-like properties: smooth, localized place fields, phase precession and considerable trial-by-trial variability. Quotation withheld because this record is not cleared for redistribution. feature-tuned field potentials (FFPs) feature-tuned field potentials (FFPs) Quotation withheld because this record is not cleared for redistribution. theta oscillation theta oscillation Quotation withheld because this record is not cleared for redistribution. place cells place cells Quotation withheld because this record is not cleared for redistribution. feature-tuned field potentials (FFPs) feature-tuned field potentials (FFPs) Quotation withheld because this record is not cleared for redistribution. action potentials action potentials Quotation withheld because this record is not cleared for redistribution. local field potentials local field potentials Quotation withheld because this record is not cleared for redistribution. position of the rat position of the rat Quotation withheld because this record is not cleared for redistribution. trial-by-trial variability of neuronal activity trial-by-trial variability of neuronal activity Quotation withheld because this record is not cleared for redistribution. 6.7 ± 0.2 cm (LFP) and 9.2 ± 0.2 cm (spiking) — Median position-decoding error from field potentials versus from spikes Quotation withheld because this record is not cleared for redistribution. 53%, 69%, and 79% — Share of total signal variance carried by the position-tuned components in each of three rats Quotation withheld because this record is not cleared for redistribution. 8 to 9 Hz — Frequency of the hippocampal theta oscillation that carries the position code Quotation withheld because this record is not cleared for redistribution. 4 to 80 Hz — Band of the broadband LFP analysed with sparse coding Quotation withheld because this record is not cleared for redistribution. ~50 — Number of FFPs recovered from two 256-electrode arrays that together covered the whole T maze Quotation withheld because this record is not cleared for redistribution. 32-, 64-, or 256-site — Channel counts of the silicon probes implanted in the hippocampus Quotation withheld because this record is not cleared for redistribution. 10-s — Delay period in the waiting area during the T-maze spatial alternation task Quotation withheld because this record is not cleared for redistribution. 6.7 ± 0.2 cm (LFP) vs 9.2 ± 0.2 cm (spiking) — median position decoding error, the paper's headline comparison Quotation withheld because this record is not cleared for redistribution. 8 to 9 Hz — frequency of the hippocampal theta rhythm treated as the carrier Quotation withheld because this record is not cleared for redistribution. 53%, 69%, 79% — share of total LFP variance carried by the position-tuned components, per rat Quotation withheld because this record is not cleared for redistribution. ~50 FFPs — number of position-tuned components recovered from two 256-electrode arrays in the T maze Quotation withheld because this record is not cleared for redistribution. 32-, 64-, or 256-site probes — recording density 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.1250444.