Zusammenfassung
Compared to brain-computer interfaces (BCI) based on electroencephalography [1], a BCI based on functional magnetic resonance imaging (fMRI) allows to record non-invasively the
activity of the whole brain with high spatial resolution. This allows to feed back local brain
activity to learn voluntary regulation of a region of interest (ROI). Using real-time fMRI, brain
responses were visualized and provided to the subject with a delay of less than 1.5 seconds
[2]. As on-line feedback, the difference of the mean BOLD signal of two regions of interest
approximating supplementary motor area (SMA) and parahippocampal place area (PPA) was
presented [3]. Eight volunteers learned to successfully control the differential neurofeedback
signal. While regulating activity in SMA and PPA, three subjects performed behavioural tasks
which are known to be characteristic for these regions. Preliminary data suggest that subject's
performance systematically changed depending on the self-regulated level of activity in SMA
and PPA. During up/down regulation of SMA and PPA, subjects had to carry out an externally
triggered bimanual nger sequence. Reaction times, i.e. onset of the movement, correlated
negatively with activity in the SMA. Hence, the more the subjects increased activity in the SMA
the faster the movement was executed. To test behavioural consequences of regulating the PPA
we made use of the fact that it is involved in encoding information for memory. Wagner et al [4]
have shown that the higher the activity of the PPA during encoding of visually presented words,
the more likely the words were recognized in a memory test. We presented 106 words while
subjects regulated activity of SMA and PPA. Interestingly, all subjects remembered the words
better in the SMA-up/PPA-down condition than in the SMA-down/PPA-up condition. Possibly,
concurring activity in the PPA might occupy resources and, thereby, interfere with other tasks
utilizing the same region. Taken together, our data suggest that self-regulation of circumscribed
functional areas leads to characteristic behavioural effects. This offers the opportunity to study
behaviour dependent on self-regulated brain activity.
