Information from Lay-Language Summaries is Embargoed Until the Conclusion of the Scientific Presentation

260—Striate Cortex: Plasticity

Sunday, November 10, 2013, 1:00 pm - 5:00 pm

260.01: Role of sleep spindle in early visual areas and perceptual learning

Location: Halls B-H

">*J. BANG, T. WATANABE, Y. SASAKI;
Brown Univ., Providence, RI

Abstract Body: Sleep is known to consolidate memory. However, the underlying neural mechanism has yet to be completely understood. Here, to better understand the mechanism for the facilitatory effect of sleep on visual perceptual learning, we measured fine-scaled spatio-temporal neural activity during sleep after training of texture discrimination task (TDT) using a multimodal neuroimaging technique that combines MEG and MRI. A leading hypothesis suggests that the sleep spindle activity during the non-rapid eye movement sleep (NREM) is involved in the facilitatory effect. Since the TDT is associated with changes in the region of the early visual areas that retinotopically corresponds to the trained visual field quadrant, we tested whether the strength of sleep spindles in early visual areas is correlated with the facilitatory effect during sleep on TDT.
Young and healthy participants underwent an MRI session after 3 nightly MEG sessions including 1 adaptation, pre-, and post-training sleep. Before the post-training sleep, TDT was conducted twice; the initial test and intensive training in one visual field quadrant. After the post-training sleep, we conducted a re-test of TDT. Wavelet-transformed MEG during sleep was combined with high-resolution MRI to constrain the current locations to the cortical mantle individually. We measured the strength of the sigma activity that represents sleep spindles in the 2 cortical parts in early visual areas which retinotopically corresponds to the trained and untrained visual field quadrants, based on the retinotopic mapping. The results showed that sigma activity yielded greater power increase in the trained region of early visual areas than in the untrained region of early visual areas after training. In addition, the power change of sigma activity was significantly correlated with the learning amount in TDT. This suggests the feasibility of the hypothesis that the sleep spindle activity is involved in the consolidation of TDT during sleep.

Lay Language Summary: Sleep is known to boost visual perceptual learning but its underlying neural mechanisms are not yet fully understood. Here we found that sleep spindles, which are characteristic brain waves during non-rapid eye movement (NREM) sleep, play a key role in boosting up visual perceptual learning, in the trained part of early visual areas, at least in the first sleep cycle.
It is important to note that we used a new non-invasive multi-modal neuroimaging technique to measure the neural correlates involving learning facilitatory action during sleep in human subjects. Electroencephalography (EEG) and magnetoencephalography (MEG) have a fine temporal resolution in the order of milliseconds but relatively poor spatial resolution. In contrast, magnetic resonance imaging (MRI) has a fine spatial resolution in the order of millimeters, but relatively coarse temporal resolution. In our study, we combined temporal information of the brain from EEG and MEG, and spatial information of the brain from MRI to capture brain activation in the order of milliseconds and millimeters.
While the topic of how sleep enhances learning has been old, the underlying neuronal processes have not been completely elucidated. It is controversial which spontaneous neural oscillations are involved in the learning facilitatory process during sleep. Among various neuronal activities, sleep spindles and slow-waves have been proposed as key mechanisms for consolidation. Experimentally, it was shown that neural stimulation with frequency that mimics sleep spindles (11-14 Hz) induce long-term potentiation (LTP). Thus, it has been proposed that newly acquired memories are reactivated via sleep spindles. On the other hand, slow-waves (0.5-2 Hz) can induce long-term depression (LTD), which may lead to synaptic downscaling. This study attempted to investigate which oscillation is involved in the consolidation of visual perceptual learning, by using a multimodal neuroimaging technique that combines MEG, EEG and MRI. This combination technique allowed us to obtain information about brain activation with high spatio-temporal resolution.
9 subjects underwent an MRI session after 3 nightly MEG sessions including 1 adaptation, pre-, and post-training sleep. Before the post-training sleep, texture discrimination task (TDT), which is well-known to induce location-specific learning within the trained visual field was conducted in one visual field quadrant. After the post-training sleep, a re-test of TDT was conducted. The brain activity during sleep was recorded using MEG and polysomnography (PSG). We combined MEG data with anatomical brain structures to source-localize the spontaneous brain oscillations to individual cortical space. By using a retinotopic mapping technique, we localized two different regions of early visual areas, which retinotopically correspond to trained and untrained visual fields. Subsequently, a systemic frequency analysis was used to obtain power for oscillations from these identified cortical regions. We selectively chose delta and sigma bands, which correspond to the frequency of slow-waves and sleep spindles. Then, we investigated which oscillation activity was involved in the consolidation during NREM sleep.
Within the sleep stage 2, but not within the slow-wave sleep stage, we found that the power of sigma activity within trained region of early visual areas was higher in the post-training sleep compared to pre-training sleep. The power increase of sigma activity was greater in the trained region compared to the untrained region of early visual areas. Interestingly, the difference of power increase between trained and untrained region of early visual areas was positively correlated with the performance improvement.
These results suggest that sleep spindles are involved in the consolidation of visual perceptual learning. Since sleep spindles would induce LTP in cortical synapses, it is plausible that the trained visual representation used in our study might have been reactivated via sleep spindles during sleep stage 2.