The very same thing: Extending the object token concept to incorporate causal constraints on individual identity

Abstract

The contributions of feature recognition, object categorization, and recollection of episodic memories to the re-identification of a perceived object as the very same thing encountered in a previous perceptual episode are well understood in terms of both cognitive-behavioral phenomenology and neurofunctional implementation. Human beings do not, however, rely solely on features and context to re-identify individuals; in the presence of featural change and similarly-featured distractors, people routinely employ causal constraints to establish object identities. Based on available cognitive and neurofunctional data, the standard object-token based model of individual re-identification is extended to incorporate the construction of unobserved and hence fictive causal histories (FCHs) of observed objects by the pre-motor action planning system. It is suggested that functional deficits in the construction of FCHs are associated with clinical outcomes in both autism spectrum disorders and later-stage stage Alzheimer's disease.

Keywords: Alzheimer’s disease; action planning; autism spectrum disorders; binding; episodic memory; medial temporal cortex; posterior parietal cortex.

Figure 1.

Components of a categorization-based model of individual re-identification. A. Binding of an occurrent object file representing a novel individual results in the encoding of three distinct representations, all of which capture the occurrent features and motion of the novel individual: a timestamped episodic memory representing the event in which the novel individual is participating, a timestamped object token representing the occurrent state of the novel individual, and a new “singular” category. B. Binding of an occurrent object file representing a familiar individual results in the encoding of a timestamped episodic memory representing the event and a timestamped object token representing the occurrent state of the individual. The singular category representing the individual may be updated to incorporate altered features, or may accumulate exemplars depending on the details of the model. The notions t1 and t2 represent timestamps.

Figure 2.

Incorporation of fictive causal histories into a categorization-based model. A. Binding of occurrent object and event files to a previous (timestamped t1) object token, associated episodic memory and associated singular category generates a new object token (timestamped t2) linked to the previous object token by an FCH and extrapolated forward by a projected future. Both interpolation and extrapolation are based on the motion and action constraints available in the singular category as it enters the binding process, that is, the motion and action constraints it encoded as of t1. B. Binding an occurrent object and event file to existing linked lists of object tokens and associated episodic memories appends current (timestamped tn+1) object tokens and episodic memories to the linked lists and updates the feature and motion information in the singular category as required by the current object and event files.

Figure 3.

Schematic representation of the temporal-parietal activation loop proposed by the BIC-FCH model. HC = hippocampus. IPL = inferior parietal lobule. LFG = lateral area of the fusiform gyrus. MFG = medial areas of the fusiform gyrus. MTG = medial temporal gyrus. PHC = parahippocampal cortex. PRC = perirhinal cortex. SPL = superior parietal lobule. STS = superior temporal sulcus. TPJ = temporal-parietal junction.