The Effects of Sleep on Neuronal Coding in Cortical Layers and Behavioral Performance

Brief periods of rest have a beneficial effect on cognitive and behavioral performance in a wide variety of tasks involving multiple modalities, including visual stimuli [6, 7, 8, 9, 3]. In 1924, Jenkins and Dallenbach first tested whether sleep improved performance yet we do not fully understand how [10]. The major limitation in our understanding of the function of rest has been the lack of an adequate animal model to allow us to study how rest influences brain networks in real time. Research on the benefits of rest and, more broadly, sleep has been exclusively dominated by invasive neurophysiological investigations in small mammals [11] (e.g. mice, rats, or cats) and non-invasive (fMRI, EEG) investigations in humans [8]. However, the behavioral repertoire of small mammals is relatively limited and importantly, they do not share human sleep cycles as they are primarily nocturnal with polyphasic, fragmented daytime sleep patterns [8, 12]. Additionally, it is currently impossible to perform highly invasive neurophysiological experimentation in humans to understand the mechanisms of the function of rest at the single cell or network level. To overcome these limitations, we plan to study the function of rest in nonhuman primates, as they offer several key evolutionary advantages, such as a sleep cycle very similar to that of humans [13, 14] and the ability to easily learn complex tasks [15].

In this proposal, we will test the influence of sleep (20-minutes (min) of rest) on neuronal coding across cortical layers in mid-level visual cortex (V4) and on perceptual performance in non-human primates. We will focus on the effects of short-term sleep (20-min of rest) rather than full-cycle sleep for the following reasons: (i) brief naps or rest periods (sleep stages 1 and 2) at the time scale proposed here have been demonstrated to consolidate memory and improve behavioral and perceptual performance [16, 17], (ii) our preliminary data showed an increase in low frequency LFP activity during rest, which was scored as sleep stages 1 and 2, and (iii) the visual discrimination task proposed here requires that our experiments be performed in a head-fixed and restrained body condition where it is easier to train monkeys to rest for 20 minutes than induce longer sleep. As to the optimal brain area to study sleep, among all sensory cortical areas, the visual cortex is the best understood in terms of receptive field properties and circuitry [18]. Thus, it provides a unique opportunity for investigating the impact of sleep on neural network coding and perceptual performance. Specifically, area V4 is hierarchically closer to decision-making areas and lesion studies have suggested that V4 plays a key role in perceptual learning [19]. The anatomical circuity underlying V4 function (similar to the rest of the neocortex) is composed of repeated motifs of laminar circuits, which are organized into six layers??“supragranular (SG; layers I-III), granular (G; layer IV), and infragranular (IG; layers V-VI). V4 sends feedforward projections through SG to areas involved in perceptual decisions and feedback projections through IG to primary visual cortex [14]. Individual neurons in V4 are strongly correlated to behavior [20] and thus specifically the output layers of V4, SG and IG, which transmit information to the rest of the brain, are likely to show stronger links with perception.

In this proposal, we will for the first time, test the layer-specific effects of sleep in the neocortex. We hypothesize that rest enhances the accuracy of population coding in visual laminar circuits by desynchronizing neural network activity. One possible explanation for desynchronous activity post-rest is the synchronous activity during rest is likely to cause the depression, or downscaling, of local intracortical synapses [21, 22]. However, synaptic depression is an asymmetric process ??“ excitatory synapses will be downscaled more than inhibitory ones [23, 24]. Consequently, asymmetric synaptic depression will cause a shift in the balance of local circuits towards inhibition after rest. Since inhibition closely follows excitation to reduce co-fluctuations in cortical responses, and thus reduce correlated variability (even when inhibition increases only modestly), local inhibition has been proposed as a mechanism for decorrelated responses in cortical networks [4, 25]. Overall, our findings will complement current theories of the function of sleep that state that sleep provides the optimal setting to (i) allow for synaptic descaling to maintain synaptic homeostasis [21], (ii) consolidate memories via neuronal re-activation [26], and (iii) remove neurotoxic waste products [27]. Our study will be the first to conclusively address the first two theories, as previous studies were unable to test the interplay of sleep, network coding, and post-sleep perceptual performance. The results of this proposal will be of importance to theoreticians and physiologists, as they will expand our understanding on how sleep modulates cortical states and cognition.

An estimated 50-70 million Americans suffer from a chronic sleep disorder [28]. Although daytime naps provide symptomatic relief [29, 30], they do not address the underlying disorder. The proposed research can serve as the basis for devising interventions. Specifically, understanding coding variability is essential in implementing neural prosthetics. Likewise, this research can help design stimulation protocols targeting visual cortices of patients that suffer from reduced alertness in visual perception due to chronic insomnia. Overall, the findings will help us understand the effect of sleep on cortical dynamics and lay the groundwork to build interventions in millions of patients that suffer from disorders of sleep.

Did you like this example?