Dependence of coherence measurements on EEG derivation type

The impact is reported of different EEG derivation types on short-term changes in the inter-hemispheric coherence between the left and right sensorimotor areas, during the planning and execution of right index finger movements. Data are recorded during an event-related paradigm in which cued index finger movements are made: Event-related coherence analysis is then applied to the monopolar (nose reference) data, as well as different reference-independent derivations such as bipolar, local average reference and source derivation. The results show that inter-hemispheric coherence between sensorimotor areas is dependent on the EEG derivation type. An increase in coherence during movement is found with nose reference and bipolar data, whereas for local average reference and source derivations, low inter-hemispheric coherence is observed, with no change in the coherence during movement. It is concluded that the coherence increase seen with nose reference data is due to an indirect effect of mu rhythm desynchronisation, rather than any increase in synchrony of the mu rhythms themselves. Local average reference and source derivations better reflect the activity of the underlying cortical structures (the mu generating networks), and coherence analysis using these derivations shows that the mu rhythms of left and right hemispheres are not coherent.