Neural Mechanisms for Auditory Memory Sequencing and Prediction
Lead Research Organisation:
Newcastle University
Department Name: Biosciences Institute
Abstract
Human and nonhuman animals rely on memories of the past to anticipate the future, weighted by the reality of the present in our ever-changing sensory world. As universal as it is for neural systems to integrate information from the past and present to forecast the future-and the striking impact on quality of life when these functions fail-fundamental empirical questions remain on how sensory, memory and predictive signalling interactions occur from neuronal levels to systems.
Substantial scientific progress has been made in studying hippocampal- and non-hippocampal-dependent learning and memory. In parallel, both the prefrontal cortex and hippocampus have now been implicated in non-classical (statistical) learning of sequential sensory dependencies separated in time, a building block for human language and intellect. Moreover, both regions relay 'top-down predictive' signals to the sensory cortex that interface with veridical input from the senses. Integrative theoretical frameworks have emerged, raising important questions that require direct empirical evidence on how neural systems interact during sensory, memory (retrospective) and predictive (prospective) functions. A scientific roadblock has been the paucity of approaches combining memory and prediction tasks, research focusing on one or a few key nodes in the system, and the lack of parallel studies in an animal model and humans using the same task.
We propose a timely cross-species neurophysiological approach during an auditory delayed sensory predictions task. Both monkeys and humans can learn the task via statistical learning in tens of minutes, and we will leverage the capability for simultaneous neuronal recordings from the prefrontal cortex, hippocampus and auditory cortex in macaque monkeys (Aim 1) and humans (Aim 2). The task harnesses statistical learning that can induce sensory sequence learning and predictions within minutes, using probabilistically variable sound sequences that manipulate "what" has and will happen and "when" (the delay between the sounds).
The cross-species approach will provide new knowledge into the interactions of a neural system integrating functions vital for daily life, and how these compare between the monkey brain and the human brain.
Substantial scientific progress has been made in studying hippocampal- and non-hippocampal-dependent learning and memory. In parallel, both the prefrontal cortex and hippocampus have now been implicated in non-classical (statistical) learning of sequential sensory dependencies separated in time, a building block for human language and intellect. Moreover, both regions relay 'top-down predictive' signals to the sensory cortex that interface with veridical input from the senses. Integrative theoretical frameworks have emerged, raising important questions that require direct empirical evidence on how neural systems interact during sensory, memory (retrospective) and predictive (prospective) functions. A scientific roadblock has been the paucity of approaches combining memory and prediction tasks, research focusing on one or a few key nodes in the system, and the lack of parallel studies in an animal model and humans using the same task.
We propose a timely cross-species neurophysiological approach during an auditory delayed sensory predictions task. Both monkeys and humans can learn the task via statistical learning in tens of minutes, and we will leverage the capability for simultaneous neuronal recordings from the prefrontal cortex, hippocampus and auditory cortex in macaque monkeys (Aim 1) and humans (Aim 2). The task harnesses statistical learning that can induce sensory sequence learning and predictions within minutes, using probabilistically variable sound sequences that manipulate "what" has and will happen and "when" (the delay between the sounds).
The cross-species approach will provide new knowledge into the interactions of a neural system integrating functions vital for daily life, and how these compare between the monkey brain and the human brain.
Technical Summary
A fundamental aspect of cognition and language is creating internal models of the sensory world that can be used to replay past events and simulate what the future may hold. As universal as it is for neural systems to actively integrate sequential information from the sensory world with internal memories and forecasts, the significant impact when these functions fail is evident. However, how neuronal systems mechanistically achieve this integration remains an open question.
A scientific roadblock has been the paucity of approaches combining memory and prediction tasks, research focusing on one or a few key nodes in the system, and the lack of parallel studies in an animal model and humans using the same task. To break through this roadblock, we propose a timely cross-species neurophysiological approach during an auditory delayed sensory predictions task. Both monkeys and humans can learn the task via statistical learning in tens of minutes, and we will leverage the capability for simultaneous neuronal recordings from prefrontal cortex, hippocampus and auditory cortex in macaque monkeys (Aim 1) and humans (Aim 2).
We plan to achieve the two objectives as follows:
1) Characterise fronto-temporal neuronal and neural system interactions in macaque monkeys conducting an auditory delayed sensory prediction task. We will examine neuronal (single unit) responses and local-field potentials (LFPs) from simultaneous array recordings of prefrontal, hippocampal and auditory cortical sites in macaque monkeys during the task.
2) Conduct parallel study of human neurosurgery patient intracranial recordings during the delayed sensory predictions task. Single unit and LFP intracranial recordings will be obtained from neurosurgery patient participants being evaluated for medically refractory epilepsy. These data will be collected from clinical electrodes within prefrontal, hippocampal and auditory cortical regions during the same task administered in Aim 1.
A scientific roadblock has been the paucity of approaches combining memory and prediction tasks, research focusing on one or a few key nodes in the system, and the lack of parallel studies in an animal model and humans using the same task. To break through this roadblock, we propose a timely cross-species neurophysiological approach during an auditory delayed sensory predictions task. Both monkeys and humans can learn the task via statistical learning in tens of minutes, and we will leverage the capability for simultaneous neuronal recordings from prefrontal cortex, hippocampus and auditory cortex in macaque monkeys (Aim 1) and humans (Aim 2).
We plan to achieve the two objectives as follows:
1) Characterise fronto-temporal neuronal and neural system interactions in macaque monkeys conducting an auditory delayed sensory prediction task. We will examine neuronal (single unit) responses and local-field potentials (LFPs) from simultaneous array recordings of prefrontal, hippocampal and auditory cortical sites in macaque monkeys during the task.
2) Conduct parallel study of human neurosurgery patient intracranial recordings during the delayed sensory predictions task. Single unit and LFP intracranial recordings will be obtained from neurosurgery patient participants being evaluated for medically refractory epilepsy. These data will be collected from clinical electrodes within prefrontal, hippocampal and auditory cortical regions during the same task administered in Aim 1.
