Periodic Reporting for period 3 - RememberEx (Human Subcortical-Cortical Circuit Dynamics for Remembering the Exceptional)
Reporting period: 2022-05-01 to 2023-10-31
There is a special place in our memory system for events that “stick-out”. We remember better those events that violate predictions generated by the prevailing context, particularly because of surprise or emotional impact. What are the human subcortical and cortical circuits that are involved in upregulating memory for exceptional events? Answering this question will help understand how we form and retrieve long-term memories for important life events and is a critical step for combating pathologies associated with memory dysfunction, including neurodegenerative diseases such as Alzheimer’s, and psychiatric conditions like schizophrenia. In the case of dementia, the rapidly growing incidence of this condition brings with it growing socio-economic burden. Human lesion and non-invasive functional imaging data, motivated by findings from animal models, have identified subcortical structures, deep in the brain, that are critical for upregulating hippocampal function during salient event memory. However, mechanistic understanding of these processes in humans remains scarce, and requires better experimental approaches such as direct intracranial recordings from, and focal electrical stimulation of, these subcortical structures. The objective of this project is to characterise human subcortico-cortical neuronal circuit dynamics associated with enhanced episodic memory for salient stimuli by studying direct recordings from human hippocampus, amygdala, nucleus accumbens, ventral midbrain and cortex. Within this framework, we are defining the electrophysiological mechanisms underlying amygdala-hippocampal-cortical coupling that lead to better memory for emotional stimuli. We are extending the hippocampal role in detecting unpredicted stimuli to define its role in orchestrating cortical dynamics in unpredictable contexts, and aim to discover the neuronal response profile of the human mesolimbic dopamine system during salient stimulus encoding. The predicted results will offer several conceptual breakthroughs, particularly regarding hippocampal function and the role of dopaminergic ventral midbrain and temporal pole in memory. The knowledge gained from this project is a fundamental requirement for designing therapeutic interventions for patients with memory deficits and other neuropsychiatric disorders.
We have made several key discoveries regarding the human brain circuitry that mediates memory enhancement for salient events. At the centre of this circuit is the hippocampus. Using intracranial recordings, we have characterised the hippocampal electrophysiological responses to unexpected stimuli, and shown how these differ along the hippocampal long axis. The broader relevance is that projections to other brain areas differ along this axis, indicating that the hippocampus may be broadcasting qualitatively different information to different brain regions in response to breaches of expectancy. The upregulation of memory for unexpected events also involves the mesolimbic dopamine system. By studying patients undergoing deep brain stimulation to the nucleus accumbens, a component of this system, we have identified an optimal stimulation site that leads to memory enhancement, particularly for infrequent events, with potential clinical applications in patients with memory impairments.
We experience breaches of expectation in everyday life, like encountering a black swan, and we also find ourselves in contexts that are more or less predictable. Calculating the predictability of a context is useful for memory formation, as this prior knowledge can optimise the balance between top-down vs bottom-up processing in the brain. We have discovered that unpredictable contexts are associated with an increase in the rate of hippocampal sharp wave–ripples (SWRs), events associated with highly synchronous neural firing. These SWRs drive a suppression of ongoing gamma activity in early visual areas, which we interpret as the hippocampus preparing the visual system for unpredictable inputs.
Memories for emotional events are among the strongest we carry with us. According to the long-standing modulation hypothesis, enhanced memory for emotional events results from a modulation of hippocampal activity by the amygdala. We have elucidated how this modulation occurs. During encoding of emotional pictures, human amygdala theta oscillations organise hippocampal gamma activity via phase amplitude coupling. Critically, we found that the phase of amygdala theta to which the hippocampal gamma is coupled determines subsequent remembering 24h later. Strikingly, this phase difference corresponded to a time interval that enabled lagged coherence between fast gamma oscillations in both amygdala and hippocampus.
We experience breaches of expectation in everyday life, like encountering a black swan, and we also find ourselves in contexts that are more or less predictable. Calculating the predictability of a context is useful for memory formation, as this prior knowledge can optimise the balance between top-down vs bottom-up processing in the brain. We have discovered that unpredictable contexts are associated with an increase in the rate of hippocampal sharp wave–ripples (SWRs), events associated with highly synchronous neural firing. These SWRs drive a suppression of ongoing gamma activity in early visual areas, which we interpret as the hippocampus preparing the visual system for unpredictable inputs.
Memories for emotional events are among the strongest we carry with us. According to the long-standing modulation hypothesis, enhanced memory for emotional events results from a modulation of hippocampal activity by the amygdala. We have elucidated how this modulation occurs. During encoding of emotional pictures, human amygdala theta oscillations organise hippocampal gamma activity via phase amplitude coupling. Critically, we found that the phase of amygdala theta to which the hippocampal gamma is coupled determines subsequent remembering 24h later. Strikingly, this phase difference corresponded to a time interval that enabled lagged coherence between fast gamma oscillations in both amygdala and hippocampus.
Our knowledge of the human brain circuitry that mediates memory formation, and memory enhancement for salient events, largely derives from non-invasive brain imaging studies. However, these lack either the temporal or spatial resolution to characterise the local electrophysiological responses in structures deep in the brain, and how different structures interact to form memories for surprising or emotional events. In RememberEx, we have overcome these limitations be recording from rare patients with electrodes implanted in deep structures. The insights provided by this approach have been striking, ascribing specific electrophysiological mechanisms to detection of unexpected events, processing of unpredictable contexts, and forming memories for emotionally aversive events. Furthermore, we have defined a brain circuit associated with voluntary action-induced memory enhancement, and localised a memory “sweet spot” in the nucleus accumbens where deep brain stimulation leads to memory enhancement, particularly for unexpected stimuli.
In the remainder of the project, we will build on these results to delve deeper into the circuit-level dynamics of memory formation for salient events. Having worked out how the amygdala and hippocampus couple to form emotional memories, we will now study the contribution of a nearby cortical area, the temporal pole, in this process. Rare patients with focal lesions in this area show profound deficits in emotional memory, and ongoing neuropsychological, MRI and electrophysiological studies in these patients will characterise the nature of this deficit and how these lesions may alter activity in distant brain regions. Single unit recordings will be analysed to 1) probe whether dissociable medial temporal responses to low vs high spatial frequency faces are evident at the single neuron level, and 2) determine whether individual neuron(s) show a stereotypical response during sharp wave ripples before and after stimuli presented in an unpredictable context. We will use state of the art magnetoencephalographic recordings in patients undergoing deep brain stimulation to probe how stimulation of the memory “sweet spot” in the nucleus accumbens affects hippocampal responses during encoding of unexpected events. Lastly, intraoperative electrical recordings during deep brain stimulation surgery will provide insights into the responses of the mesolimbic dopamine system during the encoding of unexpected events.
In the remainder of the project, we will build on these results to delve deeper into the circuit-level dynamics of memory formation for salient events. Having worked out how the amygdala and hippocampus couple to form emotional memories, we will now study the contribution of a nearby cortical area, the temporal pole, in this process. Rare patients with focal lesions in this area show profound deficits in emotional memory, and ongoing neuropsychological, MRI and electrophysiological studies in these patients will characterise the nature of this deficit and how these lesions may alter activity in distant brain regions. Single unit recordings will be analysed to 1) probe whether dissociable medial temporal responses to low vs high spatial frequency faces are evident at the single neuron level, and 2) determine whether individual neuron(s) show a stereotypical response during sharp wave ripples before and after stimuli presented in an unpredictable context. We will use state of the art magnetoencephalographic recordings in patients undergoing deep brain stimulation to probe how stimulation of the memory “sweet spot” in the nucleus accumbens affects hippocampal responses during encoding of unexpected events. Lastly, intraoperative electrical recordings during deep brain stimulation surgery will provide insights into the responses of the mesolimbic dopamine system during the encoding of unexpected events.