Abstract
Abstract
Episodic memory, as the name suggests, is the repository for data in the brain relating to what happens second by second that enables the individual to reconstruct episodes and events from the past, covering timescales from not long after birth until a few seconds ago. If an individual experiences stress, however, the operation of the autonomous stress response system may have collateral effects on the encoding and/or consolidation of the associated memory, possibly leading to memory pathology. Clinical observations show qualitatively similar effects on memory for stressors over a range of magnitudes, from those sufficiently minor to have no long-term effects to those causing major trauma with life-changing consequences, although the experimental literature is divided on whether stress enhances or impairs memory overall. Although memory pathology is an accepted symptom of stress, the underlying causal mechanisms within the brain are not generally understood. The purpose of the work reported in this thesis is to resolve the ambiguity in the literature on the effects of stress on memory and to increase understanding of the neurological mechanisms behind observed pathologies. The ambiguity concerning the effects of stress on memory was investigated experimentally (within-groups, repeated measures over 2 days and between groups, experimental group = 73, control group = 66), using a novel application of the trauma film paradigm. The findings demonstrated that the same stressor may result in both memory enhancement and impairment depending on individual differences among the participants, and on the transient operation of the autonomous stress response relative to the timing of elements within the stress-inducing episode. This led to the hypothesis that the initial stress response and the processes restoring homeostasis respectively enhance and inhibit memory processes. The experimental timings confirmed that the operation of the autonomous system must be through neurotransmitter synthesis (assumed to be noradrenaline) and its consequent action on the neural circuits supporting the memory engram. Because there are no readily available tools to test the hypotheses arising from the experimental programme at the level of neural circuits, a novel computer network model was reverse-engineered from the known features of episodic memory and the connectivity between the brain structures recruited to achieve them and implemented using the NEURON neuroscience software. The model was shown to be able to support all the processes required for episodic memory and implied that encoding and consolidation are not sequential but are initiated simultaneously. The network supporting the initial memory engram was identified with Papez’ (1937) circuit and further identified with the proposed episodic buffer element of Baddeley and Hitch’s (1974) working memory model, probably the first time that any element of that model
has been identified with a specific neural substrate. The network was shown to undergo circuit-wide Hebbian (1949) potentiation through a process termed “racetrack potentiation”, which is not otherwise described in the literature, and which operates in the theta-wave frequency range supporting gamma-wave frequency oscillations, a combination thought to be behind many emergent processes in the brain. Finally, consideration of simultaneity led to the conclusion that the human experience of “now” may be controlled by G-protein parameters, specifically the decay time. Extension of the computer model to the effects of stress demonstrated the credibility of the assumption of noradrenaline synthesis modulating glutamate processes by binding to G-protein-coupled metabotropic receptors. Initial increased efficiency of glutamate processes was shown to enhance memory encoding and consolidation, and the overshoot of noradrenaline and glutamate reabsorption to inhibit these same processes with the opposite effect on memory. These two mechanisms correspond to Solomon and Corbit’s (1974) a-process and b-process. With a suitable choice of time constants, application of the extended model to a distribution representing individual differences in the experimental and control group populations allowed a quantitative reproduction of the timings of the experimental results. The synthesis of the computational and experimental results gives strong support for the hypothesis that the initial effects of stress enhance memory encoding and consolidation through release of noradrenaline and its subsequent modulation of glutamate processes, and that the processes restoring homeostasis through glutamate and noradrenaline reabsorption inhibit them, thereby simultaneously resolving the ambiguity in the literature and identifying the neurological mechanisms underlying the effects of stress on episodic memory.