The physical basis of alpha waves in the electroencephalogram and the origin of the "Berger effect".

Synchronised activity, differing in phase in different populations of neurons, plays an important role in existing theories on the function of brain oscillations (e.g., temporal correlation hypothesis). A prerequisite for this synchronisation is that stimuli are capable of affecting (resetting) the phase of brain oscillations. Such a change in the phase of brain waves is also assumed to underlie the "Berger effect": when observers open their eyes, the amplitude of EEG oscillations in the alpha band (8-13 Hz) decreases significantly. This finding is usually thought to involve a desynchronisation of activity in different neurons. For functional interpretations of brain oscillations in the visual system, it therefore seems to be crucial to find out whether or not the phase of brain oscillations can be affected by visual stimuli. To answer this question, we investigated whether alpha waves are generated by a linear or a nonlinear mechanism. If the mechanism is linear - in contrast to nonlinear ones - phases cannot be reset by a stimulus. It is shown that alpha-wave activity in the EEG comprises both linear and nonlinear components. The generation of alpha waves basically is a linear process and flash-evoked potentials are superimposed on ongoing alpha waves without resetting their phase. One nonlinear component is due to light adaptation, which contributes to the Berger effect. The results call into question theories about brain-wave function based on temporal correlation or event-related desynchronisation.