Generalization and Specificity of Visual Learning During Sleep

The proposed study plans to collect pilot data for a study that investigates the role of sleep in the formation of feature and location specificity of VPL. Ten participants will be asked to do a visual task and nap inside the magnetic resonance imaging scanner with polysomnography.

Investigators will collect data from 10 participants, which was determined by the power analysis based on the previously published data.

Investigators will require each subject to attend six visits for six sessions. Session 1 includes the pretest of orientation detection, Session 2 orientation decoder construction, Session 3 adaptation nap, Session 4 baseline nap, Session 5 orientation detection training stage + posttraining nap, and Session 6 posttest of orientation detection. To assess the extent of performance improvement resulting from visual training, investigators will conduct behavioral sessions using an orientation detection task before and after a training session (Sessions 1, 5, and 6). For decoding orientations in each region of interest (ROI), investigators will construct an orientation decoder (Session 2) to utilize in later stages (Sessions 4 and 5) to estimate how much each orientation is represented in each ROI during sleep. Since it is well documented that young and healthy human subjects with no chronic sleep problems do not sleep well in the very first session of a sleep experiment, known as the first night effect (FNE), the adaptation nap session (Session 3) will be conducted so that participants will familiarize themselves with the experimental setting, to minimize the FNE and facilitate normal sleep from the next sleep session on. Each session will be done on separate days. In Session 5, the posttraining nap will follow the detection training on the same day.

As a behavioral measure, investigators will use an orientation detection task in the left or right visual field. Investigators will randomly determine the trained stimulus location for each participant. The stimulus will be masked with noise. Investigators will estimate the threshold signal-to-noise ratio (SNR) for an orientation detection task. Participants will engage in a two-interval forced choice task for orientation detection: they will be asked to indicate which interval (the first or the second) contained stripes by pressing either the '1' or '2' button on a keyboard while maintaining their gaze on a fixation point at the center of the display throughout trials. Investigators will present several orientations, and one of them will be specifically trained for the training session. Training usually decreases the threshold SNR, meaning that a participant could detect the trained orientation with more noise, while the threshold SNR would remain the same for untrained orientations.

An orientation decoder is a statistical algorithm that allow us to estimate how much feature information is there in a local brain region. Investigators will apply the decoder to each ROI, including early visual areas and the frontal regions, while participants are asleep to estimate whether the trained orientation is intensively processed.