Closed-loop brain stimulation augments fear extinction in male rats

Abstract

Dysregulated fear reactions can result from maladaptive processing of trauma-related memories. In post-traumatic stress disorder (PTSD) and other psychiatric disorders, dysfunctional extinction learning prevents discretization of trauma-related memory engrams and generalizes fear responses. Although PTSD may be viewed as a memory-based disorder, no approved treatments target pathological fear memory processing. Hippocampal sharp wave-ripples (SWRs) and concurrent neocortical oscillations are scaffolds to consolidate contextual memory, but their role during fear processing remains poorly understood. Here, we show that closed-loop, SWR triggered neuromodulation of the medial forebrain bundle (MFB) can enhance fear extinction consolidation in male rats. The modified fear memories became resistant to induced recall (i.e., 'renewal' and 'reinstatement') and did not reemerge spontaneously. These effects were mediated by D2 receptor signaling-induced synaptic remodeling in the basolateral amygdala. Our results demonstrate that SWR-triggered closed-loop stimulation of the MFB reward system enhances extinction of fearful memories and reducing fear expression across different contexts and preventing excessive and persistent fear responses. These findings highlight the potential of neuromodulation to augment extinction learning and provide a new avenue to develop treatments for anxiety disorders.

Fig. 1. Closed-loop SWR-timed medial-forebrain bundle electrical stimulation attenuates fear memories.

a Schematics of the experimental design. b A custom threshold crossing algorithm was used to trigger the MFB stimulation following online detections of SWRs. c Closed- loop stimulation consisted of MFB stimulation during the detected SWR events, open-loop stimulation was similar to closed-loop but stimulation was jittered from SWRs (top). Representative LFP signals from dorsal hippocampus showing SWR events and stimulation pattern (right). d Average online detection rate of SWR events (PD Positive detection; FPD False positive detection, MD Missed detection; n = 8) (one-way ANOVA: F (2,21) = 602.1, P < 0.0001). e Delay of stimulation triggering from the beginning of the SWRs. The largest number of stimuli (blue peak) were delivered between 18 and 21 ms after the SWR onset (black line: time zero). f No difference in sleep architecture between closed-loop (CL) and non-stimulated (NS) animals during non-REM (NREM) sleep (Unpaired t test: t (22) = 0.5977, P = 0.5561, two-tailed) and REM sleep (Unpaired t test: t (22) = 0.9459, P = 0.3545, two-tailed). Data represent mean ± SEM (Number of sessions: (NS) n = 13; (CL) n = 11). g No difference in fear expression in response to the CS+ following training between the three experimental groups (Kruskal–Wallis test: H = 1.737, P = 0.4195) (non-stimulated (NS) n = 9; open-loop (OL) n = 9; closed-loop (CL) n = 8). h No difference between the fear expression of the three groups during the first 5 CS+ block after first and last extinction days. There was a significant decrease in fear expression over time (mixed ANOVA: F (1,23) = 164.2, P < 0.0001, time factor). Values are normalized to the freezing expressed immediately after footshock training (i.e., “Test”). i Animals exposed to closed-loop stimulation required less extinction sessions to achieve the remission criterion compared to the open-loop and non-stimulated groups (Kruskal–Wallis test: H = 13.60, P = 0.0011). j Closed-loop neuromodulation-induced lower fear expression during the renewal test in a hybrid context (Kruskal–Wallis test: H = 16.21, P = 0.0003). k Closed-loop neuromodulation prevented spontaneous fear recovery 25 days after extinction (Kruskal–Wallis test: H = 10.38, P = 0.0056). l Closed-loop neuromodulation produces the greatest reduction in fear between post-training and remote testing following extinction (Kruskal–Wallis test: H = 13.06, P = 0.0015). *P < 0.05, **P < 0.01, ***P < 0.001. Bar plots and error bars represent medians and interquartile ranges, individual data points are also displayed. Detailed statistics are shown in Supplementary Data 1. Source data provided as a Source Data file. Silhouettes on a, b are obtained from https://github.com/eackermann/ratpack under MIT License.

Fig. 2. Contribution of fear extinction and side effects on co-storaged memories during closed-loop MFB stimulation.

a Schematics of the experimental design. Fear conditioning and test was performed as before. Closed-loop animals were exposed to 3 consecutive SWR-triggered stimulation sessions without extinction. No difference was found in fear expression in response to the CS+ following training (Mann–Whitney test: U = 7, P = 0.3095, two-tailed) b and renewal (Mann–Whitney test: U = 6, P = 0.2222, two-tailed) c between the groups (non-stimulated (NS) n = 5; closed-loop (CL) n = 5). d Before fear conditioning, animals were trained in a visual cue forced alternation T-maze task until achieving 80% of correct choice. Next, animals were exposed to fear conditioning, extinction and stimulation following Fig. 1. e T-maze performance was unaltered during the experiments regardless of the stimulation type (Unpaired t test: P > 0.05 on all instances, two-tailed) (open-loop (OL) n = 6; closed-loop (CL) n = 6). Individual performance of the animals is shown for open-loop f and closed-loop g. Bar plots and error bars represent medians and interquartile ranges, individual data points are also displayed. Detailed statistics are shown in Supplementary Data 1. Source data provided as a Source Data file. Silhouettes on a and d are obtained from https://github.com/eackermann/ratpack under MIT License.

Fig. 3. SWRs are required for the extinction of fear memories.

a The behavioral protocol was performed as before, but SWR-triggered VHC stimulation was performed for 1 h following each extinction session. b Representative LFP signals from dorsal hippocampus showing intact and disrupted SWR events. c No difference in fear expression in response to the CS+ following training between the two experimental groups (Mann–Whitney test: U = 14, P = 0.2977, two-tailed; open-loop (OL) n = 9; closed-loop (CL) n = 5). d No difference between the fear expression of the two experimental groups during the first 5 CS+ block from first and last extinction day. However, there was a significant decrease in fear expression over time (mixed ANOVA: F (1,12) = 65.00, P < 0.0001, time factor) e SWR-disrupted animals require more days to achieve the extinction criterion (Mann–Whitney test: U = 6, P = 0.0280, two-tailed). f SWR-disrupted animals show high fear expression during renewal (Mann–Whitney test: U = 4, P = 0.0120, two-tailed). g No difference in fear expression during reinstatement (Mann–Whitney test: U = 20, P = 0.7972, two-tailed). *P < 0.05, ***P < 0.001. The bar plots and error bars represent medians and interquartile ranges, and individual data points are also displayed. Detailed statistics are shown in Supplementary Data 1. Source data provided as a Source Data file. Silhouettes on a are obtained from https://github.com/eackermann/ratpack under MIT License.

Fig. 4. The closed-loop neuromodulation-induced enhancement of extinction is mediated by Rac1 and D2Rs in the BLA.

a The behavioral protocol and closed-loop neuromodulation were performed as before and immediately after each extinction session, the BLA was bilaterally microinfused with the Rac1 inhibitor NSC2376, D1R antagonist SCH23390, or D2R antagonist sulpiride. b The locations of the cannula tips in each animal are shown, with colors representing the different experimental groups. c No significant difference in fear expression was observed in response to the CS+ following training between the four experimental groups (Kruskal–Wallis test: H = 5.430, P = 0.1429) (closed loop (CL + Vehicle) n = 6; closed-loop + NSC2376 (CL + RAC1 ANT) n = 5; closed-loop + SCH23390 (CL + D1 ANT) n = 5; closed-loop + sulpiride (CL + D2 ANT) n = 5). d No difference between the fear expression of the four experimental groups during the first 5 CS+ block from first and last extinction day. However, there was a significant decrease in fear expression over time (mixed ANOVA: F (1,34) = 175.1, P < 0.0001, time factor). e NSC2376 and sulpiride injected animals required more extinction sessions to achieve the extinction criterion (Kruskal–Wallis test: H = 16.16, P = 0.0011). f Sulpiride suppress the extinction enhancement induced by closed-loop neuromodulation during renewal (Kruskal–Wallis test: H = 14.84, P = 0.0020). g Animals treated with NSC2376 and sulpiride exhibited fear recovery compared to animals injected with vehicle (Kruskal–Wallis test: H = 12.55, P = 0.0057). Bar plots and error bars represent medians and interquartile ranges, individual data points are also displayed. Detailed statistics are shown in Supplementary Data 1. Source data provided as a Source Data file. Silhouettes on a and b are obtained from https://github.com/eackermann/ratpack under MIT License.