The occipital EEG of subjects differing in reported vividness of mental imagery was monitored during a series of mental imagery tasks. (i) Vivid imagers (selected on the basis of the Betts Vividness of Imagery Scale) had a higher mean dominant alpha frequency (m.d.f.) than weak imagers (P < 0·025), but there were no individual differences for alpha abundance. The difference for m.d.f. held for only one task - an eyes-open, minimal imaging condition, (ii) Greater alpha suppression followed presentation of High Imagery (High I) words than followed presentation of Low Imagery (Low I) words (P < 0·005). (iii) High I words elicited more imagery than Low I words, for both vivid and weak imagers (P < 0·005). (iv) Reported elicited imagery was greater for the vivid imagers (P < 0·005). (v) For two voluntary imaging tasks, alpha suppression was greater in that task reported by the subjects as being harder to visualize (P < 0·01), thus reversing the findings for (ii) above, (vi) Subjects completed the Eysenck Personality Inventory. None of the above findings could be attributed to individual differences in either extraversion or neuroticism. It is concluded that the act of imaging per se does influence the EEG and that alpha suppression is not merely a by-product of the arousing nature of imaging tasks. However, difficulty in the act of imaging also influences the EEG. The equivocal results of earlier research may therefore be accounted for in terms of the differential effects of instruction to image, upon the EEG, as demonstrated in the present study.