Although previous research indicates that sleep architecture is largely intact in primary insomnia (PI), the spectral content of the sleeping electroencephalographic trace and measures of brain metabolism suggest that individuals with PI are physiologically more aroused than good sleepers. Such observations imply that individuals with PI may not experience the full deactivation of sensory and cognitive processing, resulting in reduced filtering of external sensory information during sleep. To test this hypothesis, gating of sensory information during sleep was tested in participants with primary insomnia (n = 18) and good sleepers (n = 20). Sensory gating was operationally defined as (i) the difference in magnitude of evoked response potentials elicited by pairs of clicks presented during Wake and Stage II sleep, and (ii) the number of K complexes evoked by the same auditory stimulus. During wake the groups did not differ in magnitude of sensory gating. During sleep, sensory gating of the N350 component was attenuated and completely diminished in participants with insomnia. P450, which occurred only during sleep, was strongly gated in good sleepers, and less so in participants with insomnia. Additionally, participants with insomnia showed no stimulus-related increase in K complexes. Thus, PI is potentially associated with impaired capacity to filter out external sensory information, especially during sleep. The potential of using stimulus-evoked K complexes as a biomarker for primary insomnia is discussed.