Direct modulation of aberrant brain network connectivity through real-time NeuroFeedback

Abstract

The existence of abnormal connectivity patterns between resting state networks in neuropsychiatric disorders, including Autism Spectrum Disorder (ASD), has been well established. Traditional treatment methods in ASD are limited, and do not address the aberrant network structure. Using real-time fMRI neurofeedback, we directly trained 3 brain nodes in participants with ASD, in which the aberrant connectivity has been shown to correlate with symptom severity. 17 ASD participants and 10 control participants were scanned over multiple sessions (123 sessions in total). Desired network connectivity patterns were reinforced in realtime, without participants' awareness of the training taking place. This training regimen produced large, significant long-term changes in correlations at the network level, and whole brain analysis revealed that the greatest changes were focused on the areas being trained. These changes were not found in the control group. Moreover, changes in ASD resting state connectivity following the training were correlated to changes in behavior, suggesting that neurofeedback can be used to directly alter complex, clinically relevant network connectivity patterns. Significance StatementMany disorders are characterized by underlying abnormalities in network connectivity. These abnormalities are difficult to address with explicit training procedures (which are unlikely to target the specific abnormalities). Covert neurofeedback however, can directly target these networks, positively reinforcing the desired connections. We have developed a method for reinforcing correlations in real-time, and show that such training is effective, inducing significant, long-lasting changes in connectivity between aberrant networks in Autism SpectrumDisorder. This provides a potential mechanism for modulating aberrant correlation structures in other clinical groups as well.

Overt social interaction and resting state in young adult males with autism: core and contextual neural features

“…This pattern departs qualitatively from the results reported by some previous data-driven studies of whole-brain functional connectivity in male adolescents and adults, including those by our own lab (e.g. Gotts et al, 2012;Ramot et al, 2017; see also Anderson et al, 2011;Cerliani et al, 2015;Cheng et al, 2015;Di Martino et al, 2014). To examine this potential discrepancy, we therefore compared the patterns of task-based and resting-state functional connectivity using the regions showing a main effect of Group in the current study.…”

Neurofeedback: Principles, appraisal, and outstanding issues

“…Conversely, what sort of target can people learn to attain? For instance, while some evidence suggests that people can learn to modulate brain connectivity (Koush et al., ; Ramot, Grossman, Friedman, & Malach, ; Ramot et al., ; Ruiz et al., ) or even particular transient spatially extended patterns (Diaz Hernandez et al., ), and real‐time (Watanabe et al., ) as well as temporally non‐local properties of brain activity (Ros et al., ) there is as yet no clear indications as to the range of feasible NF targets, and questions such as whether or not the measure of brain activity needs to be unique ought to be addressed in earnest. One may conjecture that almost any target may be attainable, so long as it can be expressed in a convenient way, though ultimately the essential ingredient is that cognitive control be able to move brain dynamics to the desired functional regime.…”

Connectome-wide search for functional connectivity locus associated with pathological rumination as a target for real-time fMRI neurofeedback intervention

“…To design an optimal neurofeedback signal for the identified FC associated with an RRS score, we performed a simulation to calculate an online real-time FC feedback signal. Here, two methods of online connectivity neurofeedback signal, sliding-window correlation (Gembris et al, 2000) and the two-point algorithm (Ramot et al, 2017), were evaluated. The sliding-window correlation is a z-transformed Pearson correlation between ROIs within a time window.…”

“…The feedback signal is calculated as a binary value; e.g., when a participant is trained to increase the connectivity, positive feedback (+1) is given if the two regions have the same change direction (e.g., an increase or decrease); otherwise, no feedback (0) is given. The original introduction of the two-point method (Ramot et al, 2017) used a control ROI to cancel a signal change unspecific to the target connectivity. With the control ROI, positive feedback is given when the two target regions have the same change direction as well as that is different from the direction in the control region.…”

“…The simulation analysis was performed for the functional connectivity identified in the connectome-wide analysis described above. Two measures of online functional connectivity, sliding-window correlation and the two-point algorithm (Ramot et al, 2017), were evaluated. Through a data-driven search for the precise location of connectivity associated with rumination severity and optimization of a connectivitybased neurofeedback signal, we introduced and demonstrated a data-driven, process-based framework of designing a rtfMRI-nf treatment protocol with precise targeting and an optimally designed neurofeedback signal.…”