Retrieval of spatial representation on network level in hippocampal CA3 accompanied by overexpression and mixture of stored network patterns

Retrieval of stored network activity pattern has been shown as a competitive transition from one attractor state to another, orchestrated by local theta oscillation. However, the fine nature of this process that is considered as substrate of memory recall is not clear. We found that hippocampal network recall is characterized by hyperactivity in the CA3 place cell population, associated with an “overexpression” of the retrieved network pattern. The overexpression was based on recruitment of cells from the same (recalled) spatial representation with low expected firing probability at the given position. We propose that increased place cell activation during state transitions might facilitate pattern completion towards the retrieved network state and stabilize its expression in the network. Furthermore, we observed frequent mixing of both activity patterns at the temporal level of a single theta cycle. On a sub-theta cycle scale, we found signs of segregation that might correspond to a gamma oscillation patterning, as well as occasional mixing at intervals of less than 5 milliseconds. Such short timescale coactivity might induce plasticity mechanisms, leading to associations across the two originally decorrelated network activity states.


Tetrode positions
The tetrodes were slowly lowered in CA3 within 2-3 weeks after the surgery while the rat was resting in a comfortable pot outside of the recording chamber. The recording reference electrode was positioned in corpus callosum. Additional reference for EEG was placed in stratum lacunosum moleculare.

Recording procedures
Neural activity was recorded while the rat was exposed to a procedure described by Jezek et al. (2011, bellow). The signal was recorded differentially against the reference tetrode. The hyperdrive was connected to a multichannel, impedance matching, unity gain headstage and its output was conducted through a 82-channel commutator to a Neuralynx digital 64 channel data acquisition system.
The signal was band-pass filtered at 600 Hz-6 kHz. Unit waveforms above individually set thresholds (45-70 uV) were time-stamped and digitized at 32 kHz. The position of the light emitting diodes on the headstage was tracked at 50 Hz to assess the animal's spatial coordinates. For the purpose of this study only data from intervals when the rat's movement speed exceeded 5 cm/sec were used. Broadband EEG from each tetrode was recorded continuously at 2000 Hz.

Spike sorting and cell classification
Spikes were sorted manually using 3D graphical cluster-cutting software (SpikeSort, Neuralynx). The feature space consisted of three-dimensional projections of multidimensional waveform amplitudes and energies. The putative pyramidal units were classified as a place cell if their mean firing rate exceeded 0.1 Hz and their firing rate maps displayed coherence >0.6 and sparsity <0.4 for at least one template session. Autocorrelation and crosscorrelation functions were used as additional separation tools. Cells with mean firing frequency above 10 Hz and with the peak to through duration shorter than 0.3 milliseconds were classified as interneurons.

Behavioural apparatus and training
In brief, rats were trained to set distinct representations of two environments of the identical shape (60x60 cm, 50 cm high walls) that differed in visual cues on their walls and floor, respectively. Black curtains surrounded the apparatus to prevent the animal to see the rest of the room. In environment A the LEDs organized in a circle were placed bellow the translucent floor and an additional light was placed on one of the walls to polarize the box. Box B was illuminated by a 60-cm-long array of LEDs lining 40 cm of the upper edge of the wall opposite to the directional LED in A and 20cm of one of the adjacent walls. The same amount of LED units was used across both environments. There were no other lights available except the LEDs in the apparatus. The training had four stages (Suppl. Fig. 1).
Initially, the boxes were located next to each other, connected by a corridor (20x20cm, width x length) that allowed the rat to freely move between them at least in three trials (20 min each) in order to develop a different set of path integrator coordinates for each box. Trials were separated by 20 min breaks while the rat rested on a towel in a pedestal outside of the curtains (phase 1). Whereas the initial phase 1 was repeated for three subsequent days, the other phases took one day each. In phase 2, the passageway was removed and the animal explored the boxes individually on alternating trials (3 trials in each, 10 min breaks). The next day (phase 3), the boxes were replaced by another box of the same size, made of the same material and equipped with both sets of lights. Its position was box was alternating at the two original locations between the trials. When presented at the original position of box A, the set of LEDs defining A was lighted; when in the place of box B, the respective lights of B were switched on instead. Again, there were three trials in each environment on alternating occasions. In the final phase, the box was presented at a central location between the two original box positions. Again, the animal received two pairs of three alternating trials with each set of lights for two consecutive days. Across all stages, at the beginning of each trial, the rat was placed into the environment with the eyes gently covered by the experimenter's palm without any further disorientation. Between trials, the boxes were thoroughly cleaned and dried. On the test day, the rat was first exposed to each box configuration (A and B, respectively) for two 10 min reference recordings. In the third trial, the lights identifying given environment (for example, A) were switched instantaneously to the other configuration (for example, B). Such 'teleportations' were performed every subsequent 40-60 s for 10 min. Another two reference sessions were recorded afterwards. The animal rested for 10 min as usual between the trials. The experiment was repeated on the subsequent day after lowering the electrodes deeper into the CA3 if sufficient amplitudes from new cells could be recorded. Animal movement was motivated by small crumbles of cookies thrown into the box at 10-20 s intervals. Three flavors of food were used: vanilla, chocolate and cereals without any additional flavor. Vanilla and unflavored were offered in one light configuration, chocolate and unflavored in the other in a balanced way. During teleportation trials, only unflavored food was offered.

Histology
After the experiment was finished, the rat was overdosed with Equithesin and was perfused intracardially with saline followed by 4 % formaldehyde. Brain coronal sections (30 μm) were stained with cresyl violet. Traces of all 14 tetrode locations were identified. Each tip location was considered as the place in the section before the tissue damage became negligible. Only recordings from tetrodes with their tips in CA3 were used in this study. The rats were trained to explore two environments, each of iden cal box shape but defined by a specific configura on of light cues. a) During Phase 1, the rats moved freely between the boxes connected with a corridor to support forma on of orthogonal place cell maps for each context. In Phase 2 the corridor was closed but both boxes retained their posi ons. In Phase 3 the rats explored a single box (equipped with both sets of light cues) with alterna ng context iden ty at the original loca ons. During the test phase, the box was placed in the middle between the original box loca ons. b) On the test day, the rats explored ini ally both contexts as in the final stage of the training (the reference sessions), followed by a teleporta on session. The teleporta on session started with light cues corresponding to one of the contexts and a er 40-60 s of explora on the lights were instantaneously switched to the alterna ve configura on. The procedure was repeated each 40-60 s throughout the session. The teleporta on session was followed by two other reference sessions, one for each of the environments.  Supplementary figure 4 a) An example distribu on of environment specificity index (ESI) values for one recording day. The ac ve CA3 place cells tend to show firing preference for one of the environments. b) Performance of ESI decoder on data from reference sessions per each recording day. Theta bins with detected expression of either map were classified based on their correspondence to the actual environment iden ty. The ver cal axis represents the instances where during reference in context A, the map A was correctly decoded (true A = 1 -false B), the horizontal axis corresponds to instances where map A was decoded in context B (false A = 1 -true B). c) Distribu on of ESI values within 20 theta cycles before and a er the cue switch. The posi ve values correspond to specificity for the current environment. The post-teleporta on distribu on displays marked increase in ac vity specific for the alterna ve context. d) Distribu on of decoded network states detected during the reference epochs of the teleporta on session and the post-teleporta on periods. During the 20 theta bins long post-teleporta on period, there is a robust increase in the 'incorrect' and in the 'mixed' states, reflec ng a compe ve network dynamics. e) An example of the evolu on of popula on ac vity before and a er a teleporta on event classified according ESI. The cue switch (green bar) is followed by a transient instable period, where the network alternates between 'correct' (blue) and 'incorrect' (red) states. Considerable amount of theta bins were classified as 'mixed' (yellow), containing combined ac vity of cells highly specific for either of the environments. Grey bars represent unclassified bins with at least 2 ac ve cells, the empty slots correspond to the bins with <2 ac ve cells. f) Evolu on of coac vity (ISI <= 10 ms) incidence between cells exclusively ac ve in either of the environments during ini al template sessions. Prolonged recording a er the last teleporta on shows a long las ng coac vity formed as a response to repe ve teleporta ons (green lines).  d) The number of ac ve cells during the 'correct' and the 'incorrect' states defined by PSI values for the ac ve units. An increase in ac vity was observed during the 'correct' (3.56 ± 0.11 cells per TC reference, 4.07 ± 0.13 cells per TC post-tele, n = 153, Wilcoxon signed-rank test: z = 3.53, p = 4.1498e-04), but not during the 'incorrect' bins (3.08 ± 0.09 cells per theta cycle stable, 3.21 ± 0.13 cells per theta cycle posttele, n = 71, Wilcoxon signed-rank test: z = 0.82, p = 0.4111). e) Shi of ac vity towards the periphery of a firing field during expression of 'correct' state detected by PSI criteria (Wilcoxon signed-rank test: z=-4.20, p= 2.6477e-05). Analysis of theta oscilla on before and a er the teleporta on vent a) Example of me-frequency representa on of local field poten al (CA3 pyramidal layer) from interval -2 to +5 seconds before and a er the telepoerta on event from one experimental day (averaged across across 15 teleporta ons). Note the increase within the theta band (around 9 Hz) within the first two secons a er the teleporta on event. b) Comparison of averaged z-score (± SEM) from theta frequency bandpassed (6-11 Hz) local field poten al from intervals of two seconds before and a er the teleporta on event, respec vely. c) Averaged z-scores from pos eleporta on interval (as in c)) corresponding to an emergence of correct, incorrect and mixed popula on vectors. d) Average theta wave dura on during 2 seconds before and a er the teleporta on, respec vely.  Examples of three interneurons recorded on independent experimental days. a) Ac vity of each interneuron across consecu ve 15 teleporta on trials (from top to bo om) during the period of -20 to +50 theta cycles before/a er teleporta on. b) Normalized average ac vity per theta cycle across all teleporta on events. c) Wave-shape of each interneuron across four channels of a tetrode and the coresponding mean firing rate. d) Evolu on of averaged and normalized absolute z-score values (± SEM) of ac vity from 16 interneurons before and a er all teleporta on events in theta cycle temporal resolu on. e) Averaged absolute z-score values (± SEM) from periods of 20 theta cycles preceding and following the teleporta on events, respec vely. f) Averaged z-score values of spiking ac vity a er teleporta on (20 theta cycles) across whole sample of interneurons. Same colors correspond to simultaneously recorded units. Colors and marks also correspond to the three interneuron examples from a) to c). activity per TC % of baseline activity per TC activity per TC activity per TC