How the Awake and Sleeping Brain Tags Relevant Information for Memory
Post by Meredith McCarty
The takeaway
This research provides evidence for a mechanism by which relevant information is selected or tagged for memory consolidation in the hippocampus, a region of the brain critical for memory.
What's the science?
An essential feature of brain function is selecting relevant information to be stored in long-term memory. However, the neural mechanism by which this is accomplished has yet to be fully understood. The authors hypothesized that hippocampal neurons are involved in selecting, or “tagging”, relevant information for subsequent memory storage. This week in Science, Yang and colleagues identify a potential neural mechanism explaining how relevant information is selected for memory formation.
How did they do it?
To study this mechanism, the authors had mice perform a spatial memory task. In this task, a water reward was delivered at the left or right arm of a figure 8-shaped maze (the location of this reward alternated on each trial). The mice had to navigate through this maze and choose the correct arm (left or right) to receive the water reward (the mice were water-restricted before beginning the task). Following this memory task, mice were placed in their familiar home cage and their neural data was recorded during periods of sleep.
During the task and sleep periods, the authors recorded the spiking activity of single neurons via electrodes located in the hippocampus, measuring changes in spiking activity and identifying occurrences of sharp wave ripples (SPW-Rs). SPW-Rs are thought to be an essential feature of “offline” brain processing, during which populations of hippocampal neurons exhibit synchronized rapid firing, communicating compressed information to connected brain regions. To understand the relationship between neural activity and key decision and memory points in the task, the authors used neural decoding and analysis techniques, linking the mice’s real-time position in the maze on a given trial with the neural spiking and SPW-R recordings.
What did they find?
First, the authors found a correlation between the mouse’s position in the maze and CA1 population spiking activity and were able to successfully decode the recent locations of the mice in the maze from the population spiking activity. When considering the activity of single neurons in isolation (which exhibit representational drift, meaning dynamic variability in activity across trials despite stable behavioral responses), the decoding accuracy deteriorated. This suggests that locations and events during the trial are represented at the population level — made up of variable activity of individual neurons — as opposed to the individual level.
Next, when analyzing neural activity during the time after the mouse successfully navigated the maze and received a reward, the authors noted the emergence of SPW-R events on numerous trials. They were able to successfully decode task information from the content of these ripple events. These SPW-Rs were preceded by a decrease in theta activity, signifying the transition out of an active behavioral state. These findings suggest that SPW-Rs, theta power, and rewarding stimuli are related to memory replay and tagging events in the hippocampus.
When comparing awake and asleep neural activity, they found the occurrence of SPW-Rs during the awake task period to be a significant predictor of subsequent SPW-Rs during sleep. This suggests that the information tagged during awake SPW-Rs is subsequently replayed during sleep, furthering the process of memory consolidation.
What's the impact?
Based on these results, the authors propose a neurophysiological framework by which experiences are tagged for further memory consolidation in the hippocampus. The proposed mechanisms underlying this process are a decrease in theta power preceding the tagging of relevant information via SPW-Rs, and the subsequent repetitions of SPW-Rs during sleep that allow the replay of these events necessary for memory consolidation.
This work has implications for both experimental and clinical work, to better understand how memories are formed, and how this process is disrupted in individuals with debilitating memory impairments.
