We have wide-ranging interests in cognitive neuroscience, but three of the major themes of the research in our lab are: (a) the processing of abstract and specific information, (b) the nature of implicit memory, and (c) effects of emotion on vision and memory.
Abstract and Specific Information Processing
Humans are able to recognize shapes in the environment at several levels of abstraction. For example, imagine seeing an object on a desk. One might identify it as a “nonliving thing,” a “beverage container,” a “mug,” a “coffee mug,” or as a particular individual entity, say “my favorite MST3K-characters mug.” A fundamental question is whether visual identification of shapes takes place in a unified system capable of processing multiple levels or in at least weakly dissociable subsystems, one that identifies an abstract category of shape (e.g., “mug”) and one that identifies the specific exemplar (e.g., “my favorite MST3K-characters mug”). The question is fundamentally important; shape recognition may take place in qualitatively different ways depending on which of two subsystems is engaged.
In contrast with prominent single-system theories, we have found evidence supporting dissociable neural subsystems using familiar objects (Marsolek, 1999), unfamiliar objects (Marsolek & Burgund, 2005), word forms (Marsolek, Kosslyn, & Squire, 1992), and pseudoword forms (Burgund & Marsolek, 1997). An abstract-category subsystem operates effectively in the left cerebral hemisphere, whereas a specific-exemplar subsystem operates effectively in the right cerebral hemisphere, but stimulus and task demands modulate these hemispheric asymmetries in important ways (Marsolek, 1999; Marsolek & Burgund, 2003; Marsolek & Hudson, 1999). An abstract-category subsystem utilizes parts-based shape processing (Marsolek, 1995), whereas a specific-exemplar subsystem utilizes whole-based shape processing (Marsolek, Schacter, & Nicholas, 1996). These subsystems are differentially affected by interhemispheric transfer of visual information (Marsolek, Nicholas, & Andresen, 2002), their hemispheric asymmetries are not caused by asymmetries of other visual subsystems (Andresen & Marsolek, 2005), and they interact with non-visual subsystems in different ways (Marsolek & Andresen, 2005). These subsystems underlie differential working memory effects (Marsolek & Burgund, submitted), as well as long-term memory effects (Burgund & Marsolek, 2000). Moreover, they not only underlie different word priming effects (Marsolek, 2004), but they contribution to word recognition in different ways that help to settle a long-standing theoretical debate in reading (Deason & Marsolek, 2005; Marsolek & Deason, in press).
Much of our evidence comes from divided-visual-field studies, which allow causal inferences to be made about the neural implementations of subsystems, complementing many correlational neuroimaging techniques. Stimuli in the right (or left) visual field are projected directly to the left (or right) hemisphere, giving subsystems in that hemisphere time and stimulus-quality advantages in processing, thus finding abstract and specific effects following right and left visual-field presentations, respectively, indicates that abstract and specific processing are (at least weakly) neurally dissociated. Converging with results from experiments using this technique, other studies indicate that abstract and specific subsystems are differentially affected by the neuromodulator serotonin (Burgund, Marsolek, & Luciana, 2003) and they utilize contradictory neurocomputational processing strategies (Marsolek & Burgund, 1997). Our current work involves functional magnetic resonance imaging and event-related brain potentials to test for additional converging results.
Presently, we are excited about implications of abstract and specific visual subsystems for better understanding interpersonal social-cognitive phenomena. For example, we are finding that, when an abstract-category subsystem predominates in the perception of faces, gender stereotypes tend to be activated, whereas when both subsystems contribute to the perception of faces, gender stereotypes tend to be inhibited (Marsolek & Rothman, submitted).
Finally, we have examined abstract and specific processing in additional domains of cognition and memory. These include analogous abstract/specific distinctions in learning perceptual-motor sequences (Marsolek & Field, 1999), causal inference generation during text comprehension (Sundermeier, Virtue, Marsolek, & van den Broek, 2005), and episodic memory (Westerberg & Marsolek, 2003a). Related memory effects have been investigated in our studies of the nature of false memories (Westerberg & Marsolek, 2006; Westerberg, Steele, & Marsolek, in press). False memories can be due to inaccurate memories per se rather than misleading biases in memory posited by other researchers (Westerberg & Marsolek, 2003b), with implications for evaluating cases of recovery of repressed memories, as well as for assessing the veridicality of memories in eyewitness testimony situations.
The Nature of Implicit Memory
An important distinction in human memory is between implicit and explicit memory. Implicit memory is best exemplified by “repetition priming,” which is a facilitation in processing due to recent performance of that processing (e.g., facilitation in identifying a piano due to recently having identified a piano). This can occur independently of explicit memory, which is involves deliberate attempts to recollect previously encoded information (e.g., intentionally recollecting a particular event in your life that involved a piano). Different parts of the brain underlie the two forms of memory (e.g., Marsolek, Squire, Kosslyn, & Lulenski, 1994), but important questions remain. In particular, because it occurs outside of intention to recollect the past, the nature and purpose of implicit memory is not clear (Marsolek, 2003).
In Rethinking Implicit Memory, my colleague (Jeff Bowers) and I edited a book addressing this issue (Bowers & Marsolek, 2003). This collection provides an overarching summary of the extant perspectives on how implicit memory is conceptualized.
Most recently, we have developed a theory that accurately posits a new memory phenomenon, “antipriming,” one that provides a window into the nature of representation and learning in the brain. By this theory, the representation of a familiar shape is strengthened via small synaptic changes after the shape is identified; learning/relearning of such information is never “turned off.” This strengthening is responsible for repetition priming for that shape, but it also is responsible for antipriming of other shapes that have representations superimposed (or overlapping) with that of the primed shape. In other words, visual identification of some familiar shapes (e.g., a piano) not only enhances subsequent identification of those shapes (e.g., a piano), but also impairs the ability to identify other familiar shapes (e.g., a desk). Supporting evidence comes from experiments with young adults, amnesic patients, matched controls, and neurocomputational models (Marsolek, Schnyer, Deason, Ritchey, & Verfaellie, 2006), and our preliminary work using functional magnetic resonance imaging and event-related brain potential also supports the theory. Essentially, priming and antipriming appear to reflect ongoing adjustments of superimposed representations in neocortex. The purpose of these changes may be to learn and maintain superimposed representations, which afford useful storage and generalization benefits.
Effects of Emotion on Vision and Memory
When we recall our individual circumstances upon learning of the horrific events of the morning of 11 September 2001, many of us feel as though our explicit memories for that situation are fairly detailed, including where we were, whom we were with, and what we were doing, presumably due to the emotionality of the event. The detailed quality of such memories has led to their being termed “flashbulb memories.” However, for real world events, the accuracy with which one recalls the details of such memories cannot be known. Without an observer recording the details of the situation, the memory cannot be verified. In addition, repeated retrieval and re-thinking of the memory may be responsible for the distinctive level of detail, as opposed to the emotionality of the event. Can such putative effects of emotion on vision and memory be investigated in ways that avoid these difficulties?
We are bringing this phenomenon into the laboratory by testing whether emotional reactivity to visual scenes enhances episodic memory for the visual details in those scenes. This new method helps to overcome the methodological concerns raised in previous research done in non-laboratory settings. Our preliminary results indicate that visually-specific memories are stored for both emotional and non-emotional scenes, but certain variables affect visually-specific memory for the two kinds of stimuli in different ways (Blank & Marsolek, in preparation; Marsolek, Park, Kittur, & Kane, in preparation). These variables range from participant personality traits to patterns of eye movements that are made when initially encoding scenes.
In related work, we are testing what underlies “novelty preferences” in looking-time experiments. An invaluable method for testing memory in pre-verbal infants and non-human primates involves simultaneously presenting two visual stimuli, one that had been presented previously and one that is novel. Typically, participants look longer at the novel stimulus compared with the old stimulus. This novelty looking effect indicates some form memory, but it is not clear what form of memory is reflected. We have used emotional conditioning to discover that competitive biases in visual selection attention may be responsible for novelty looking effects, in contrast with prevailing theories that explicit memory is responsible (Snyder, Blank, & Marsolek, submitted). These results have important implications for the utility and applicability of such measures of memory in infants.
Home | Research | Publications | People | Pictures | Links | Contact