Somatosensory ECoG-based brain-machine interface with electrical stimulation on medial forebrain bundle

Abstract

Brain-machine interface (BMI) provides an alternative route for controlling an external device with one's intention. For individuals with motor-related disability, the BMI technologies can be used to replace or restore motor functions. Therefore, BMIs for movement restoration generally decode the neural activity from the motor-related brain regions. In this study, however, we designed a BMI system that uses sensory-related neural signals for BMI combined with electrical stimulation for reward. Four-channel electrocorticographic (ECoG) signals were recorded from the whisker-related somatosensory cortex of rats and converted to extract the BMI signals to control the one-dimensional movement of a dot on the screen. At the same time, we used operant conditioning with electrical stimulation on medial forebrain bundle (MFB), which provides a virtual reward to motivate the rat to move the dot towards the desired center region. The BMI task training was performed for 7 days with ECoG recording and MFB stimulation. Animals successfully learned to move the dot location to the desired position using S1BF neural activity. This study successfully demonstrated that it is feasible to utilize the neural signals from the whisker somatosensory cortex for BMI system. In addition, the MFB electrical stimulation is effective for rats to learn the behavioral task for BMI.

Keywords: Brain plasticity; Brain–machine interface; Deep brain stimulation; Somatosensory cortex; Virtual reward.

Fig. 1

Schematic representation of the BMI system operation. The ECoG signals of the S1BF were collected from a freely moving rat in real-time. The power spectrum of the gamma band (40–70 Hz) in the right (green) and left (blue) barrel cortex were compared and transformed into a dot position on a screen. Simultaneously, the rat watched the screen and tried to control the dot to move it to the screen center. MFB electrical stimulation was provided as a reward for task completion

Fig. 2

Electrode implantation and position confirmation in the rat brain a image acquired during surgery depicting the position of electrodes and anchors. b Locations of the implanted electrodes. Dashed line indicates the cross section of c. c MFB (arrow) and S1BF (box) locations in the coronal brain section

Fig. 3

Intracranial self-stimulation (ICSS) a Lever presses per min by days (*p < 0.05). b Comparison of lever presses pre- and post-ICSS (*p < 0.05)

Fig. 4

A sample of transformation of neural activity to dot position during a BMI task. a Amplitude spectrum between the gamma band (40–70 Hz) of the (blue) and right barrel cortex (red). b Converted dot displacement with direction. c Cumulative position of the dot on the screen

Fig. 5

Operant conditioning BMI task. a Task success rate. (* for t-test with Day 1, p < 0.05). b Total Success Count. (* for t-test with Day 1, p < 0.05). c Time taken for a trial of the BMI task (* for t-test with Day 1, p < 0.05). d Rat in the ECoG-based BMI experiment environment

Fig. 6

Examples of success trials. The two red lines indicate a target zone displayed on the monitor. The blue dash line indicates the threshold for success. a Single success in a single trial (day 2). b Multiple successes in a single trial (day 4)