ausblenden:
Schlagwörter:
-
Zusammenfassung:
The necessary condition for sensory information to enter consciousness remains an open question. Studies on humans and non-human primates indicate the need for reverberant neural activity lasting for 10s-100s of milliseconds. Properties of this reverberant activity, especially the involvement of early sensory areas, are heavily debated. Here, we addressed these issues using behavior and causal circuit manipulation in mice. Using visual backward masking, we first tested whether visual perception in mice requires reverberant neural activity. Mice were trained to discriminate between locations (± 45 deg eccentricity) of a briefly presented target (grating, 16 ms duration, 10 contrast). After reaching threshold discriminability (d’ > 2), bilateral masks (plaids, 16 ms, 100 contrast) were introduced at various delays to target onset. Mice performed at chance up to stimulus onset asynchronies (SOAs) of 66.7 ms. Together with the results of rats (Watanabe et al. SfN 2014), the capacity of the mask to render the target invisible beyond the time of target offset indicates the necessity of reverberant activity for rodent visual perception. Next, to investigate whether V1 takes part in the crucial circuitry for perception, we replaced the visual mask with optogenetic suppression of V1. Unlike the visual mask eliciting activity throughout the visual system, optogenetic suppression acts locally and is ideal for testing the necessity of activity in distinct processing stages. We trained mice with ChR2 expression targeted at V1 PV+ inhibitory interneurons in the position discrimination task, in which we replaced the visual mask by bilateral optogenetic suppression of V1 activity. We induced suppression by activation of inhibitory interneurons (1.5 s duration) at various delays. Behavioral performance was at chance when V1 was suppressed before the onset of target evoked activity, but was significantly above chance when suppressed later (> 32 ms). Since only the initial transient V1 response, and not the later sustained V1 activity, is required for perception, V1 seems to function as an initial supplier of visual information to the reverberant loop, but does not play a crucial role in its maintenance. Finally we investigated visual forward masking where a visual mask (16 ms, 100 contrast) precedes the target (16 ms, 10 contrast). Behavioral results showed extended SOA ranges of invisibility (16-150 ms). Interestingly, neural responses to high contrast masks (16 ms duration) resulted in a prolonged neural suppression effect (150 ms) that matched the SOAs of invisibility. The results indicate that forward masking blocks relaying of visual information as early as area V1.