A Computational Model of Perceptual and Mnemonic Deficits in Medial Temporal Lobe Amnesia

Abstract

Damage to the medial temporal lobe (MTL) has long been known to impair declarative memory, and recent evidence suggests that it also impairs visual perception. A theory termed the representational-hierarchical account explains such impairments by assuming that MTL stores conjunctive representations of items and events, and that individuals with MTL damage must rely upon representations of simple visual features in posterior visual cortex, which are inadequate to support memory and perception under certain circumstances. One recent study of visual discrimination behavior revealed a surprising antiperceptual learning effect in MTL-damaged individuals: With exposure to a set of visual stimuli, discrimination performance worsened rather than improved [Barense, M. D., Groen, I. I. A., Lee, A. C. H., Yeung, L. K., Brady, S. M., Gregori, M., et al. Intact memory for irrelevant information impairs perception in amnesia. Neuron, 75, 157-167, 2012]. We extend the representational-hierarchical account to explain this paradox by assuming that difficult visual discriminations are performed by comparing the relative "representational tunedness"-or familiarity-of the to-be-discriminated items. Exposure to a set of highly similar stimuli entails repeated presentation of simple visual features, eventually rendering all feature representations maximally and, thus, equally familiar; hence, they are inutile for solving the task. Discrimination performance in patients with MTL lesions is therefore impaired by stimulus exposure. Because the unique conjunctions represented in MTL do not occur repeatedly, healthy individuals are shielded from this perceptual interference. We simulate this mechanism with a neural network previously used to explain recognition memory, thereby providing a model that accounts for both mnemonic and perceptual deficits caused by MTL damage with a unified architecture and mechanism.

Recognition memory shielded from semantic but not perceptual interference in normal aging

“…One theory of the role of MTL structures in cognition -the Representational-Hierarchical (R-H) account Cowell, Bussey, & Saksida, 2006;Sadil & Cowell, 2017)has begun to amass empirical support for its explanation of mnemonic and perceptual deficits following brain damage (Barense et al, 2005;Barense, Gaffan, & Graham, 2007;Barense, Groen, et al, 2012;Bartko, Cowell, Winters, Bussey, & Saksida, 2010;Bartko, Winters, Cowell, Saksida, & Bussey, 2007a, 2007bLee, Buckley, et al, 2005;Lee, Bussey, et al, 2005;Lee, Levi, Davies, Hodges, & Graham, 2007;McTighe, Cowell, Winters, Bussey, & Saksida, 2010). However, to our knowledge this theory has not been applied to the effects of aging on memory (but see Scheerer & Marrone, 2014).…”

A Roadmap for Understanding Memory: Decomposing Cognitive Processes into Operations and Representations

“…For example, the label "feature representation" implies a neural code richer in feature-based information than in information about whole objects. This can be mathematically defined, implemented in a computational model (Bussey and Saksida, 2002;Cowell et al, 2006Cowell et al, , 2010 and measured empirically (Cowell et al, 2017), and it has consequences for the outputs of that cortical region. This label thereby begins to deconstruct and specify the mechanisms of cognition.…”

“…A further advantage of representational labels is that, unlike a process such as recollection, they do not appear to explain more than they do, because they make clear what is left to be specified -namely, the operations that act upon the representations to produce a behavioral output, e.g., recall of a memory. A further advantage of operation labels (e.g., pattern completion) is that they can prompt the drawing of mechanistic parallels between two high-level phenomena, if they contribute to both (e.g., recollection and priming; see also Sadil and Cowell, 2017).…”

Visual interference can help and hinder memory: Capturing representational detail using the Validated Circular Shape Space

“…Following dissimilar interference, the reduction in retrieval success (mixture model analyses) driven by the relative increase in responses containing minimal target information (parameter-free analyses) could reflect not only the complete erasure of the target memory (Zhang & Luck, 2008;, but also other mechanisms that can render the memory temporarily inaccessible (Tulving, 1974). One possibility is that the memory trace could not be accessed due to competition from distractors (Cowell, Bussey, & Saksida, 2006;Sadil & Cowell, 2017;Schurgin et al, 2019), resulting in errors made to competitors displayed throughout the experiment (e.g., also known as "swap", "misbinding", or "temporal binding" errors; Bays, Catalao, & Husain, 2009;Chun, 1997;Diana, Peterson, & Reder, 2004;Sadil & Cowell, 2017;Schurgin, Wixted, & Brady, 2019). Moreover, participants provide their response along a 2D…”

“…In this case, swaps to similar items would manifest as reductions in precision and an increase in responses to items bearing coarse-grained similarity to the target (Bays, Catalao, & Husain, 2009;Cowell, Bussey, & Saksida, 2006;Diana, Peterson, & Reder, 2004;Sadil & Cowell, 2017;Schurgin, Wixted, & Brady, 2018). Importantly, the behavioral manifestations of these mechanisms may be effectively indistinguishable on tasks such as ours.…”