Closed-Loop Self-Neuromodulation of the Amygdala Involves Co-Modulation of the Posterior Insula

667 

Abstract Submission 

Guy Gurevitch^1,2, Naomi Fine^1,3, Ayelet Or-Borichev^1,2, Tom Fruchtman-Steinbok^1,3, Talma Hendler^1,2,3,4

¹Sagol Brain Institute, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel, ²Department of Physiology and Pharmacology, Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel, ³School of Psychological Sciences, Tel Aviv University, Tel Aviv, Israel, ⁴Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel

Guy Gurevitch  
Sagol Brain Institute, Tel Aviv Sourasky Medical Center|Department of Physiology and Pharmacology, Faculty of Medicine, Tel Aviv University
Tel Aviv, Israel|Tel Aviv, Israel

Naomi Fine  
Sagol Brain Institute, Tel Aviv Sourasky Medical Center|School of Psychological Sciences, Tel Aviv University
Tel Aviv, Israel|Tel Aviv, Israel

Ayelet Or-Borichev  
Sagol Brain Institute, Tel Aviv Sourasky Medical Center|Department of Physiology and Pharmacology, Faculty of Medicine, Tel Aviv University
Tel Aviv, Israel|Tel Aviv, Israel

Tom Fruchtman-Steinbok  
Sagol Brain Institute, Tel Aviv Sourasky Medical Center|School of Psychological Sciences, Tel Aviv University
Tel Aviv, Israel|Tel Aviv, Israel

Talma Hendler, PhD  
Sagol Brain Institute, Tel Aviv Sourasky Medical Center|Department of Physiology and Pharmacology, Faculty of Medicine, Tel Aviv University|School of Psychological Sciences, Tel Aviv University|Sagol School of Neuroscience, Tel Aviv University
Tel Aviv, Israel|Tel Aviv, Israel|Tel Aviv, Israel|Tel Aviv, Israel

Self-neuromodulation (also known as NeuroFeedback; NF) is a form of Brain Computer Interface through which individuals learn to modulate their own neural activity, by receiving reinforcing feedback regarding desired changes in neural activity patterns. Previous works have repeatedly shown individual variability in the ability to successfully modulate neural signals (Haugg et al., 2021) and the involvement of widespread brain networks other than the neural target during modulation (Emmert et al., 2016; Goldway et al., 2022). It remains an open question whether this variability stems from the direct accessibility to the target modulation or from co-modulation in other regions. To test these options, we combined data from multiple cohorts targeting Amygdala down-modulation using a validated fMRI informed EEG model (termed Electrical Finger Print; EFP, Keynan et al., 2019).

125 patients diagnosed with either post-traumatic stress disorder (n=84) or Fibromyalgia (n=41) participated in clinical studies comprising 6-10 sessions of Amygdala driven-EFP (Amyg-EFP) training (Fruchtman-Steinbok et al., 2021; Fine et al., 2023). We assessed Pre- and post-training modulation with real-time fMRI-NF. Successful modulators in the fMRI-NF were defined by a negative mean BOLD estimate during regulation compared to a passive watch period in each session. The fMRI data were preprocessed with fmriprep (Esteban et al., 2019) and further analyzed with SPM12 to demonstrate activation patterns in successful and unsuccessful modulators before and after NF training. Extra-target activation ROIs were selected and defined based on whole brain maps of the same contrast shown in previous results (Goldway et al., 2022) - See Figure 1 numbered circles - 1) Posterior insula; 2) Hippocampus; 3) Posterior cingulate cortex; 4) Medial prefrontal cortex. These activation clusters were extracted bilaterally from the first-level maps of each subjects and further assessed in a univariate ANOVA for group and session differences. Finally, a mediation analysis was used to depict the relationship between the fmri-NF target modulation and Amyg-EFP training through extra-target co-modulation.

Similar to previous works in healthy participants, we found distinct activation patterns for Amygdala down-modulation in successful compared to unsuccessful modulators across sessions. During the post-NF session relative to pre-NF, successful modulators showed enhanced distributed deactivations, while unsuccessful modulators showed enhanced distributed activations. Images were assessed for cluster-wise significance at p(FDR)<0.05; cluster defining threshold p<0.0001. (Fig 1). ROI analysis revealed a significant Group x Session x Condition interaction in the bilateral posterior insula (Fig 2A, F=9.224; p<0.012, corrected for multiple comparisons), showing that only successful modulators down-modulate this region during regulation compared to watch, and do so more after NF training with the Amyg-EFP probe (T=5.79, p<0.0001). Introducing a mediation model, we show that the association between Amyg-EFP modulation and Amygdala activation change is mediated by activation change in the posterior Insula (Fig 2B, Indirect path, Sobel Z=2.32; p<0.026).

Our results demonstrate different network patterns activated during successful and unsuccessful down-modulation of the amygdala during NF training. Examining neural changes following multi-session training emphasizes the posterior insula as a region directly involved in learning to self-modulate the amygdala. Further research may support an understanding of the causal dynamics of such successful modulation. Altogether contribution of beyond-the-target regions for NF success, could be utilized in the future for guiding personalized NF training design in neuropsychiatry.

Psychiatric (eg. Depression, Anxiety, Schizophrenia) ¹

Higher Cognitive Functions Other

Skill Learning ²

Activation (eg. BOLD task-fMRI)

BOLD fMRI

FUNCTIONAL MRI

Learning

Limbic Systems

Other - neurofeedback