The ability of human beings to keep track of the identities of objects within a visual scene has been investigated using controlled laboratory experiments for several decades. A major finding of these kinds of experiments is that how an object moves is the deciding factor in whether people judge it to remain "the same thing" over brief (a few hundred milliseconds to a few seconds) periods of continuous observation; an object that moves in the "right way" is seen as "the same thing" even if its appearance changes radically during the motion, while an object that moves in the "wrong way" is seen as a sequence of multiple different things even if its appearance does not change at all. For example, a horizontally-moving object that passes briefly behind another object and immediately re-appears on the other side is typically seen as "the same thing" even if it changes shape and color while moving. However, if a horizontally-moving object disappears behind another object and reappears on the other side at a point somewhat below where it disappeared, most people will report that one object has vanished and another, different object has appeared, even if the object's shape and color have not changed (some experiments along these lines are available from the Yale Perception and Cognition Laboratory). I have developed a neurofunctional model that describes how the dependence of object persistence judgments on observed motion could be implemented by the visual processing and visuo-motor systems of the primate brain (see "Trajectory recognition as the basis for object individuation: A functional model of object file instantiation and object token encoding"). This model depends on the recognition of visual trajectories - paths through 3-dimensional space - in a feature-independent way by specific circuits in the visuo-motor cortex. It is not currently known whether such specific trajectory-recognition circuits exist in humans; I propose several kinds of experiments that would characterize their operation if they do.

People are also able to identify individual objects as "the same thing" over longer periods (a few seconds to years) that involve extended lapses of observation; most people have no trouble identifying their spouse or their car, for example, even after not seeing them for several hours. This ability clearly cannot be explained by an ability to recognize trajectories, and standard neurocognitive models of individual identification across perceptual episodes focus on the recognition of diagnostic features that allow an object seen now to be related to memories of that object as it was seen in other contexts. Such models have difficulty accounting for the ability to re-identify objects after their features change - for example, the ability to re-identify someone who has grown a beard or dyed their hair - and cannot account for the ability to re-identify objects in the presence of other objects ("competitors") with identical or nearly-identical features. In cases involving feature change or competitors, people routinely appeal to the causal histories of objects to explain their decisions about what has remained "the same thing." If an object has not been observed for an extended period, however, its causal history cannot be known; the histories to which people appeal when making judgments about identity are, therefore, in many cases fictive. I have extended the trajectory-recognition model of within-scene identity tracking to show that the same visuo-motor systems that recognize trajectories are capable, given some assumptions that are supported by available neurofunctional data, of generating fictive causal histories of observed objects (see "The very same thing: Extending the object token concept to incorporate causal constraints on individual identity"). If this model is correct, it has the interesting consequence that people do not actually recognize objects as being the same things over extended time, but rather make up stories with which to convince themselves that they are seeing something that they have seen before. Under ordinary conditions, people make up these stories so fast that they are not conscious of doing so; they believe that they are recognizing familiar objects instantaneously, and report their experiences as experiences of recognition. People do not make up such fictive casual stories about objects voluntarily, but rather as a consequence of human neurocognitive architecture, i.e. of the "wiring" of the brain.

At the current stage of data availability, neurofunctional models of the mechanisms underlying human judgments of object identity over time are somewhat speculative. Such models are valuable, however, because they predict specific types and ranges of individual differences in object persistence judgments, and specific kinds of cognitive deficits as outcomes of either atypical development of the relevant networks in infancy or early childhood or functional disruptions of the relevant networks during adulthood. The models described above both predict that observed motions are abstracted during the encoding of memories that specify what objects participated in a given observed event, and that overly-specific encoding of motions in such memories will disrupt object re-identification abilities. Disrupted object re-identification in infancy would be expected to lead to disrupted category learning and the formation of aberrant object categories that cut across "normal" categories along dimensions defined by similarities of movement. Such aberrant category formation would, in turn, be expected to generate behavioral outcomes, including difficulties in the recognition of human individuals and language learning deficits, that are similar to those observed in autism spectrum disorders. Specific disruptions in the ability to re-identify familiar individuals are common in the later stages of Alzheimer's disease; the model of object re-identification by fictive history construction described above suggests that this amnesia for individual identity may have its roots in the visuo-motor system, not in the hippocampus as is commonly assumed.

My current work on this research track is focused on the role of the visuo-motor system in concept learning, conceptual reasoning and semantic memory, i.e. memory for facts and the meanings of words. I am particularly interested in the role of the visuo-motor system in analogical reasoning (see the system-brain project) and in encoding the "background knowledge" that human beings and other animals use to solve the frame problem, the problem of determining what has not changed following an action or an observed event.