How Sleep Improves Behavioral Performance
Post by Laura Maile
The takeaway
Sleep is known to improve learning and memory, but the underlying neural mechanisms are not well understood. In the visual and prefrontal cortex of the brain, synchrony of neural network activity is decreased following sleep, which correlates with improved performance in visual tasks.
What's the science?
Most previous research on how sleep improves learning and memory has been done in humans, where only non-invasive procedures like EEG are possible, or in rodents where studies have only investigated the influence of sleep on memory.
This week in Science, Kharas and colleagues used multielectrode arrays inserted in the brain to record neuronal activity in macaques to see how sleep affects performance on behavioral tasks. The electrodes were inserted into specific brain areas to detect the changes in activity patterns across populations of neurons during behavioral tasks and sleep, offering the authors a more accurate view of non-rapid eye movement (NREM) sleep-related activity. Using macaque monkeys allowed them to observe performance in complex tasks than is possible with other non-human mammals.
How did they do it?
Five macaque monkeys were trained to complete visual discrimination tasks while experimenters recorded their brain activity. Monkeys then completed identical tasks before and after 30-minute rest periods, where their brain activity was continually monitored. NREM sleep was confirmed during rest periods using a polysomnogram (combination of electroencephalogram, eye movement, and muscle monitoring) and video analysis of their eyes and face. Multielectrode arrays, consisting of sets of multiple electrodes inserted into distinct areas of the visual cortex and prefrontal cortex, were used to measure neural activity during the tasks and the rest period. Local field potentials (LFP) were recorded, which allowed experimenters to detect neural activity across a population of neurons and analyze patterns of synchronized activity in specific frequency bands. Next, they replaced sleep with a 30-minute period of neural stimulation in the delta frequency, using the implanted electrodes in the V4 visual cortex. They repeated the neural activity analysis and the behavioral tasks before and after stimulation. Finally, they used network modeling of the visual cortex to model the observed changes in activity seen in the visual tasks following sleep. This allowed them to determine the likely cause of the observed changes in activity and synchrony that were associated with better performance.
What did they find?
During the NREM phases of 30-minute sleep periods, LFP analysis of neuronal spiking activity showed decreased power in the gamma band and increased power in the low-frequency bands, especially the delta band, commonly associated with sleep. The increases in the delta band were associated with an increase in synchronization of neuronal firing in all tested brain areas. Performance in a visual discrimination task improved in monkeys after sleep and compared to control monkeys that sat in a dark room but were not allowed to sleep. Neural activity of the sampled population of neurons became desynchronized after sleep in all recorded areas and noise correlation decreased after sleep. Neural firing increased in all brain areas during the task after sleep but remained unchanged in monkeys that did not sleep.
Importantly, the observed changes in synchronized activity and neural firing were correlated to improved performance in behavioral tasks. They found that by stimulating V4 visual cortex in the delta frequency while the animal was awake, they were able to achieve similar effects on neural firing, synchrony of activity, and behavioral performance that occurred due to sleep. This suggests that the increase in synchronized activity seen during sleep leads to reduced synchrony during tasks completed after sleep, allowing increased accuracy of system activity and improved task performance. Finally, the authors’ network models revealed that depression of inhibitory synapses likely accounts for the observed changes in neural population activity seen during the behavioral task after sleep. This suggests that the improvements in the discrimination task may be due to a net increase in excitatory synaptic activity between cortical neurons.
What's the impact?
This study found that activity across neural networks in the visual and prefrontal cortex becomes more synchronized during sleep, but less synchronized after sleep. This decrease in synchrony is associated with increased firing and improved performance in visual discrimination tasks. This study was also the first to demonstrate successful invasive electrode stimulation of distinct brain areas to improve performance in behavioral tasks, which could lead to future improvements in brain neuromodulation in humans.