Adult Auditory Evoked Vertex Potentials In Sleep

Abstract

Slow vertex potentials, evoked by filtered clicks (1200 Hz), were recorded (C_z–M₁) from 11 adults while awake, in light sleep or in deep sleep. Sleep was induced by small doses of secobarbital, about 2.5 mg/kg.

The familiar waking response pattern, N₉₀–P₁₇₀, was judged present or absent according to 5 voltage criteria, and the percentage of correct identifications (hit rate) and the percentage of false-positive judgments were related also to the sensation levels of the stimuli. If a 17-percent false-positive rate for a single trial is considered clinically acceptable, the best voltage criterion is about 3.3 μV. This gives a hit rate of about 80% for 20 dB SL clicks.

The depth of sleep was estimated by the amount of dealta activity in the EEG, recorded with a 2 to 4-Hz pass-band, a full-wave rectifier and an integrator. Sleep response patterns, later and higher in voltage than the waking response, were clear in deep sleep (stages 3 or 4). The most characteristic and stable component was P₂₀₀–N₃₃₀, as described by Ornitz et al. [1967a and b]. For this the best voltage criterion was about 5 μV, which, for 9 of the 11 subjects, gave a hit rate of about 60% for 20 dB SL clicks.

In light sleep (stages 1 or 2) the P₂₀₀–N₃₃₀ response was often present but it could not be so reliably identified because an increased percentage of false-positive ‘responses’ required the use of a very strict criterion, about 8 μV.

Key Words:

• Electric response audiometry
• Filtered clicks
• EEG stages of sleep
• Index of delta wave activity
• Hit rate
• False-positive rate
• Evoked-response patterns
• Signal detection theory