Exploring personalized approaches for neurofeedback optimization

Neurofeedback is an intervention which is designed to alter cognitive-behavioral functions by inducing neural activity. Despite the growing evidence supporting its validity in healthy and clinical populations, there is current debate on its underlying mechanisms and its limited efficacy in mediating lasting changes to cognitive-behavioral outcomes. Recent neurofeedback studies suggest that inter-individual variations in neurophysiological and psychological factors may explain some of the heterogenous findings and may, further, inspire the individual adaptation of neurofeedback protocols. In a series of studies, we investigated how tailoring specific parameters of neurofeedback interventions may increase the overall efficacy in mediating changes to both neurophysiological and cognitive outcomes.

In study 1, we assessed the effect of self-pacing training-time during neurofeedback on cognitive outcomes and how the learning-rate of gaining control over the targeted neural feature related to changes in neurophysiological and cognitive functions measured before and after a single session of neurofeedback. Our results indicate that handing healthy participants the ability to adapt the training-time to their needs improved higher cognitive functions more than externally-pacing the training. We, further, observed that the more successful healthy participants were in regulating their brain activity, the more they increased their neurophysiological and cognitive-behavioral functions from before to after neurofeedback.

Next, we assessed person-specific neural correlates of mental rotation in study 2 by applying a machine learning approach. Modelling the time of correct responses as a function of preceding EEG activity in a mental rotation task revealed inter-individual differences in the importance of spectral and spatial EEG features for the prediction on the one hand and similar group-based EEG features as reported in previous studies on the other hand. More specifically, our findings highlighted the relevance of alpha and beta band-related activity as well as left frontal and right parieto-occipital activity for mental rotation.

In study 3, we finally investigated the efficacy of personalized target feature specification in a neurofeedback intervention by applying the methodology detailed in study 2. Our results suggest that compared to sham feedback, participants receiving personalized neurofeedback improved their mental rotation performance significantly more. Furthermore, the mediated effect of neurofeedback training was specific to mental rotation and did not transfer to other cognitive functions.

The results from the conducted neurofeedback studies provide partial support for the notion that adapting neurofeedback protocols to user characteristics increases the efficacy of neurofeedback interventions in mediating changes to neurophysiological and cognitive-behavioral functions. Future studies are required to overcome some of our limitations also related to the spatial resolution of the chosen brain imaging technique, our adaptation of the mental rotation task, and the specificity of neurofeedback as an intervention to exclusively induce changes to the targeted neural feature.