Slow and fast cortical cholinergic arousal is reduced in a mouse model of focal seizures with impaired consciousness

Abstract

Patients with focal temporal lobe seizures often experience loss of consciousness associated with cortical slow waves, like those in deep sleep. Previous work in rat models suggests that decreased subcortical arousal causes depressed cortical function during focal seizures. However, these studies were performed under light anesthesia, making it impossible to correlate conscious behavior with physiology. We show in an awake mouse model that electrically induced focal seizures in the hippocampus cause impaired behavioral responses to auditory stimuli, cortical slow waves, and reduced mean cortical high-frequency activity. Behavioral responses are related to cortical cholinergic release at two different timescales. Slow state-related decreases in acetylcholine correlate with overall impaired behavior during seizures. Fast phasic acetylcholine release is related to variable spared or impaired behavioral responses with each auditory stimulus. These findings establish a strong relationship between decreased cortical arousal and impaired consciousness in focal seizures, which may help guide future treatment.

Keywords: CP: Neuroscience; GRAB sensor; acetylcholine; auditory perception; consciousness; epilepsy; focal aware seizures; focal impaired awareness seizures; mouse model; slow waves; temporal lobe epilepsy.

Figure 1.. Induced focal limbic seizures in mice cause increased cortical slow waves and impaired behavioral responses

(A) Diagrams illustrating the experimental apparatus and the recording setup. Mice were head-fixed and allowed to run on a running wheel while simultaneously obtaining recordings from the hippocampus (HC) and the orbitofrontal cortex (OFC). An auditory stimulus indicated water was available through a lick spout. Electrical stimulation to induce HC seizures is indicated by a lightning symbol. (B) Example traces from a single experiment. The auditory stimuli are represented by vertical lines and speaker symbols. Long vertical dashed lines surrounding the lightning symbol signify the start and end of the artifact from the electrical stimulus to HC. Shorter vertical dashed lines highlight the end or start of defined periods. Seizure activity is observed in the ipsilateral (Ipsi) and contralateral (Contra) HC following the unilateral electrical stimulation in the Ipsi HC (indicated by yellow artifact). The animal stopped licking to the auditory stimulus (as indicated by a pause in the vertical blue traces) and stopped running (indicated by no change in wheel position in bottom trace) during the ictal period. (C) Zoom-in of the LFP signals (gray box in B) showing representative baseline period. (D) Zoom-in of the LFP signals (red box in B) showing representative ictal period. Note that both HCs show seizure spikes, whereas OFC shows slow oscillation activity without seizure propagation. Note that the scale is 10 times smaller for Ipsi and Contra HC in (C) than in (D). (E) Time-frequency plots showing power change from average baseline power (normalized power) at defined periods: baseline, ictal, early postictal, and late postictal periods for Ipsi HC, Contra HC and OFC. Normalized power/frequency (dB/Hz) was calculated by normalizing power spectral density relative to baseline (see STAR Methods). (F) Power spectrum resulting from the time-frequency plots. Both HC signals showed large increase of power in a broad frequency range for both HCs signals, including the delta (δ) (1–4 Hz) and beta (β) (15–30 Hz) band power (gray boxes), during the ictal period, followed by an overall decrease of power after the end of seizure. The OFC signal showed maximum increase of power in the delta range and a smaller increase in the beta range during the ictal period, followed by a progressive decrease during the early and late postictal periods. (G and H) Scatterplots showing mean delta power (G) and mean beta power (H) changes during ictal, early postictal, and late postictal periods for Ipsi HC, Contra HC, and OFC. (I) Mean lick rates aligned to sound presentation (vertical line with speaker symbol) are compared at different periods. Ictal period showed a decreased lick rate compared to baseline. Note that the lick rate before the auditory stimuli is lower during the ictal period. (J and K) Scatterplot showing average of maximum lick rate (J) and average of delay to the first lick (K) following sound presentation during each period for each recording. Seizures significantly decrease lick rate (J) and increase latency to first lick (K). (L) Scatterplot showing average wheel speed for each period for recordings where mice were running (>10 cm/s) during the baseline period (n = 171 seizures, 26 animals). Mice showed significantly decreased wheel movement during seizures. For (E)–(K), n = 225 seizures in 26 animals. Data are shown as mean ± SEM. Significance was calculated by ANOVA with Bonferroni post hoc pairwise comparisons of baseline versus each of the other time periods. **p < 0.02 and ***p < 0.01. Bsl, baseline period; Ic, ictal period; Post, early postictal period; Late, late postictal period.

Figure 2.. Decreased activity of cortical neurons during focal limbic seizures resembles slow-wave sleep

A) Schematic drawing showing general location of the unipolar multiunit activity (MUA, purple) electrode in OFC added among the bipolar local field potential (LFP) electrodes (yellow) and an example of an MUA electrode location marked by the fluorescent 1,1′’-dioctadecyl-3,3,3′’,3′’-tetramethylindocarbocyanine Perchlorate (DiI) agent in OFC after performing histology in one animal. Scale bar: 1 mm. (B) Example of a recording session including the MUA signal (purple) from the OFC. Symbols and marking of events during the recording follow the same conventions as in Figure 1. (C) Zoom-in of the tonic activity of the OFC neuronal population recorded from the MUA electrode during baseline from the gray box in (B). (D) Zoom-in of the alternating up- and down-state activity during the seizure from the red box in (B). Note that neuronal activity (indicated by gray shading) matched the OFC LFP up-state oscillations rather than HC seizure spike activity. (E) Mean time courses of MUA V_RMS change from baseline, aligned at the start of seizure, showed a decrease of neuronal population activity during the ictal period that progressively increased back to baseline during the early postictal and late postictal periods. The two full vertical black lines before the start of the ictal period in (E) correspond to a cut artifact period caused by the electrostimulation. (F) Scatterplot showing significant decrease of mean MUA V_RMS change during the ictal and early postictal periods compared to the baseline period. For (E) and (F), n = 13 seizures in 10 animals. Data are indicated as mean ± SEM. Significance was calculated with ANOVA with Bonferroni post hoc pairwise comparisons of baseline to each of the other periods. ***p < 0.001 and n.s, non-significant.

Figure 3.. Impaired seizures are associated with more cortical slow waves

(A) Example of a spared and an impaired seizure defined by the presence or absence of licking responses within a 1-s window after the auditory stimuli during the ictal period. Symbols and marking of events during the recording follow the same conventions as in Figure 1. (B) Graph showing details on spared and impaired seizures classification. A licking response is considered as a “hit” if the first lick occurred within a 1-s window after the auditory stimulus (green region), while it becomes a “miss” if the first lick happened after the 1-s window following sound or if there is no lick (pink region). A seizure is defined as “spared” if every licking responses during the ictal period is a hit (blue symbols) or is defined as “impaired” if every response during the ictal period is a miss (red symbols). (C and D) Time-frequency plots (C) and power spectra (D) showing normalized power change of OFC LFP at different periods during spared seizures versus impaired seizures. Normalized power/frequency (dB/Hz) was calculated by normalizing power spectral density relative to baseline (see STAR Methods). Note a higher increase of delta power (δ) (1–4 Hz, gray regions in D) during the impaired seizures. (E) Comparison of mean delta power changes between spared and impaired seizures during ictal, early postictal, and late postictal periods. For (B)–(E), spared, n = 72 seizures in 16 animals; impaired, n = 37 seizures in 13 animals; 23 animals in total. Data are indicated as mean ± SEM. Significance was calculated with Mann-Whitney U test for spared versus impaired seizures in each period. ***p < 0.001.

Figure 4.. Miss responses during seizures are accompanied by more cortical slow waves than hit responses

(A) Example of a recording with a seizure composed of both hit (green background) and miss (pink background) responses to the auditory stimuli during the ictal period. Symbols and marking of events during the recording follow the same conventions as in Figure 1. (B) Comparison of mean lick rates aligned to sound presentation (vertical line with speaker) at different periods for hit versus miss responses. All miss responses showed no licking responses within a 1-s window after the auditory stimuli; in hit responses, the first lick did occur within 1 s. Note that the baseline lick rate before the auditory stimuli was lower during the miss responses compared to the hit responses (see text). (C) Time-frequency plots showing normalized power change of OFC LFP in ±2-s data segments around the time of auditory stimuli (vertical line with speaker) for the hit and miss responses at different periods. Normalized power/frequency (dB/Hz) was calculated by normalizing power spectral density relative to baseline (see STAR Methods). Miss responses showed higher increase in delta power (versus baseline hit period) compared to the hit responses. (D) Comparison of the power spectrum of the ictal OFC LFP from the hit and miss responses (red box in C) showing a higher increase of delta power (gray shading, 1–4 Hz) during the miss responses than the hit. (E) Mean delta power changes of the ictal OFC LFP showing significant difference between hit versus miss responses (same data segments as C red box, and D). For (B)–(E), ictal, n = 166 hits and n = 140 misses; early postictal, n = 216 hits and n = 51 misses; late postictal, n = 220 hits and n = 51 misses during 225 seizures in 26 animals. Data are indicated as mean ± SEM Significance was calculated with Mann-Whitney U test. ***p < 0.001.

Figure 5.. Cortical ACh neurotransmission is reduced during focal limbic seizures in relation to behavior

(A) Schematic drawing showing general location of the implanted optic fiber (blue/green line) for fiber photometry recording, added among the LFP electrodes (in yellow) and an example of histology showing the tract of the optic fiber located above the region of GFP fluorescence from expression of the ACh3.0 sensor in OFC. Scale bar: 1 mm. (B) Example of a recording session including the ACh3.0 GFP signal and the isosbestic point (reference) signal. ACh3.0 signal decreased during the seizure induced by HC electrical stimulation (lightning symbol surrounding by dashed lines), while the isosbestic point signal did not show major changes. Note that the animal stopped licking (vertical blue traces, in response to sound stimuli indicated by speaker symbols and vertical lines) and running (no change in wheel position) during the seizure. (C) Mean time courses of seizure-related ACh3.0 change from behaviorally spared versus impaired seizure recordings, aligned at the start of seizure (ictal), then at start of early postictal period (postictal), and continuing into the late postictal period (late). ACh3.0 signal showed larger decrease during impaired seizures versus spared seizures. (D) Mean seizure-related ACh3.0 changes showing significantly larger reduction during the ictal and early postictal periods from impaired seizures compared to spared seizures. (E) Mean event-related ACh3.0 changes aligned to sound presentation (vertical trace with speaker) are compared between the baseline hit, ictal hit, and ictal miss responses. Two fast ACh evoked response phases are seen: a first release phase (P1, dark gray box from 0 to 250 ms) and a second release phase (P2, dark gray box from 750 ms to 1.5 s). A pre-auditory stimuli baseline period was defined from −1 to 0 s (Pre-Aud, light gray box). (F) Comparison of mean event-related ACh3.0 changes between the pre-auditory stimuli period (Pre-Aud) and the first ACh release phase (P1) for the baseline hit, ictal hit, and ictal miss responses showing significant increases of P1 versus Pre-Aud for the baseline and ictal hit responses, whereas there is no significant increase for the ictal miss responses. (G) Comparison of mean event-related ACh3.0 changes between the pre-auditory stimuli period (Pre-Aud) and the second ACh release phase (P2) for the baseline hit, ictal hit, and ictal miss responses showing significant increase of P2 versus Pre-Aud for the baseline hit responses, no significant difference for the ictal hit responses, and a significant decrease of P2 versus Pre-Aud for the ictal miss responses. For (C) and (D), spared, n = 18 seizures; impaired, n = 31 seizures in seven animals. For (E–G), n = 607 baseline hits, 128 ictal hits, 119 ictal misses during 118 seizures in seven animals. Data are indicated as mean ± SEM Significance was calculated with Mann-Whitney U test. **p < 0.02; ***p < 0.01; and n.s, not significant.