Neurofeedback Training of Alpha-band Coherence After Stroke

Background: The technology of brain-computer interfaces (BCI) enables the monitoring of brain activity and the generation of a real-time output about specific changes in activity patterns. The recorded subject receives a feedback about the neural activity associated his/her efforts and can thus learn to voluntarily modulate brain activity. There is accumulating evidence that training of motor cortex activations with brain-computer interface systems can enhance recovery in stroke patients. Here we propose a new approach which trains resting-state correlates of motor performance instead of activations related to movements. Previous studies have shown that the more resting-state alpha oscillations in the motor cortex are coherent with the rest of the brain, the better stroke patients perform in motor tasks. Furthermore, observational studies have suggested that training of alpha-band coherence in the motor cortex with neurofeedback has beneficial effects on motor performance.

Objective : This randomized controlled study aims to test the usefulness of training functional connectivity between the motor cortex and the rest of the brain with a brain-computer interface in patients with chronic stroke. We hypothesized that this network variant of neurofeedback training will lead to region and frequency specific increases in functional connectivity and to an improved function of the affected upper extremity.

Methods : 10 patients with chronic stroke and significant unilateral deficit of upper extremity motor function will perform two periods of neurofeedback training in a randomized cross-over design. In one period, they will train alpha-band coherence between intact areas around the affected motor cortex and the rest of the brain. In a control period, they will train alpha-band coherence between a control region not directly related to motor function (the medial prefrontal cortex of the healthy hemisphere) and the rest of the brain. In each period, two training sessions per week will be performed for 4 weeks. The periods are separated by at least 4 weeks. Oscillations in the brain will be reconstructed from 128 EEG channels using an adaptive spatial filter and the coherence between the target area and the rest of the brain will be calculated in real time. Coherence magnitude will be displayed in the form of a cursor on a computer screen.

Significance: This study may provide causal evidence for a role of functional connectivity in motor learning and may lead to new strategies for rehabilitation.