We aimed to estimate primary sensory evoked potential (EP) amplitude, amplitude–intensity functions and habituation in migraine patients compared with healthy control subjects and to investigate the possible relation to check size, sound and light discomfort thresholds, and the time to the next attack. Amplitudes of cortical visual evoked potentials (VEP, check size 8′ and 33′), cortical long latency auditory evoked potential (AEP N1P1; 40, 55 and 70 dB SL tones) and brainstem auditory evoked potential (BAEP wave IV–V; 40, 55 and 65 dB SL clicks) were recorded and analysed in a blind and balanced design. The difference between the response to the first and the second half of the stimulus sequence was used as a measure of habituation. Twenty-one migraine patients (16 women and five men, mean age 39.3 years, six with aura, 15 without aura) and 22 sex- and age-matched healthy control subjects were studied (18 women and four men, mean age 39.5 years). Low sound discomfort threshold correlated significantly with low levels of BAEP wave IV–V amplitude habituation (r = –0.30, P = 0.05). VEP and AEP amplitudes, habituation, and amplitude–intensity function (ASF) slopes did not differ between groups when anova main factors were considered. Control group VEP habituation was found for small check stimuli (P = 0.04), while potentiation was observed for medium sized checks (P = 0.02). The eight migraine patients who experienced headache within 24 h after the test tended to have increased BAEP wave IV–V ASF slopes (P = 0.08). This subgroup did also have a significant VEP habituation to small checks (P = 0.04). No correlation was found between different modalities. These results suggest that: (i) VEP habituation/potentiation state and brainstem activation state may depend on the attack–interval cycle in migraine; (ii) VEP habituation/potentiation may depend on spatial stimulus frequency; (iii) phonophobia (and possibly photophobia) may depend more on subcortical (brainstem) function than on cortical mechanisms; (iv) low cortical preactivation in migraine could not be confirmed; (v) EP habituation and ASF analysis may reflect sensory modality-specific, not generalized, central nervous system states in migraine and healthy control subjects.