On the electroencephalogram in man