Weten wat te doen - de neurofysiologische basis van het kortetermijngeheugen

This thesis investigates the neural network activity in the hippocampus that underlies short-term memory (STM), using spatial memory in rats as a model system. The central hypothesis is that STM is facilitated by neural representations of experiences being put into focus - or activated, potentially by repetitive neuronal activation, and recruited at critical decision-making moments. Two competing working models are evaluated: one positing that hippocampal activity while running, specifically theta oscillations in the local field potential, is sufficient on its own for STM processes (working model 1), and another suggesting that additional hippocampal activity during rest, specifically sharp wave-ripples (SWRs), is necessary for these processes (working model 2). This research aims to assess the validity of these models through a combination of behavioural tasks and in vivo electrophysiological recordings. In Chapter 2, I tested working model 2 by disrupting SWRs during various spatial STM tasks requiring rats to memorise locations over a short period. SWRs are of interest because the co-occurring replay - the firing of neurons representing a trajectory through space - has been observed to represent either past or future trajectories. Replay of past trajectories has been hypothesised to help with memorisation and replay of future trajectories with the recruitment of memories for making a memory-based decision. Surprisingly, the results indicated no significant impairment in performance following SWR disruption, suggesting that SWRs do not play a critical role in STM. Chapter 3 focused on working model 1, examining the role of theta oscillations during active navigation in facilitating STM. Utilizing high-density Neuropixel probes, I recorded large-scale hippocampal activity in rats performing an STM task in an 8-arm radial maze. Analysis revealed that the spiking activity during theta oscillations primarily represented trajectories on the current and occasionally on the upcoming maze locations, but rarely on past or alternative maze locations. This lack of deliberation over choices before visits is unexpected, as prior studies indicated that the hippocampus typically exhibits alternating representations of navigational options. Instead, these sweeps may reflect scanning of the local environment rather than explicit decision-making. I hypothesize that hippocampal activity during running supports STM by relaying spatial information to other brain regions, a notion warranting further investigation. Notably, the neural activity was the same across both sensory- and STM-based navigation phases, supporting the suggestion of the involvement of additional brain regions in mediating STM-based navigation. The results of this thesis support the first working model, emphasizing the critical role of hippocampal activity during running for STM, rather than during rest. Activation and memorization of neural representations are facilitated by temporally compressed firing during theta oscillations. While forward sweeps of potential future trajectories likely occur before navigational choices, the findings suggest these sweeps do not represent distinct task-relevant options but rather local trajectories. Memory- based deliberation for future navigations likely takes place in another brain region, such as the prefrontal cortex (PFC), which has been shown to couple with hippocampal theta oscillations during spatial STM tasks. Additionally, reactivations during SWRs are not essential for STM; instead, they may contribute to long-term memory consolidation during sleep-related reactivations.