ICOSR-Emory Cognition Meeting—Attention-worthy and Memorable
4 May 2009. The ninth and, so far, biggest, International Congress-Emory University Cognition Satellite Meeting brought schizophrenia researchers and basic neuroscientists together in San Diego 27-28 March. Our report on this meeting comes in two parts. Today’s story covers the first round of talks, chaired by Deanna Barch of Washington University in St. Louis, Missouri. The speakers began by exploring how advances in neuroscience might help us understand memory and attention deficits in schizophrenia, followed by talks about the challenges involved in translating laboratory findings into real-world interventions. Part 2 of our coverage describes the afternoon talks on the lifetime course of schizophrenia.
Three talks focused on memory for events, or episodic memory, with the first one focusing on the brain structures involved. The speaker, Charan Ranganath of the University of California at Davis, said that growing evidence hints that the medial temporal lobes tend to serve different functions, with the perirhinal cortex involved in recognizing an item as familiar, the parahippocampal cortex in remembering contextual information, and the hippocampus in tying items to their context. In the past, memory researchers have focused on the hippocampus, but the prefrontal cortex also appears to contribute to the deficits in episodic memory seen in schizophrenia. He and his colleagues recently found that five weeks of working memory training increased dopamine-1 receptor binding in the prefrontal cortex; they are currently conducting a study to see if training can improve episodic memory.
Ranganath’s colleague at the University of California at Davis, J. Daniel Ragland, addressed the contribution of the prefrontal cortex to schizophrenia-related deficits in relational memory, which involves relationships between items. Ragland thought that prefrontal dysfunction might be involved because subjects with schizophrenia and those with lesions of the prefrontal cortex perform similarly in certain ways on memory tasks. For instance, both groups show severe impairments in relational memory and have trouble generating strategies to organize information during encoding; giving them strategies improves their performance.
Past studies also suggest that the ventrolateral prefrontal cortex encodes information about specific items, whereas the dorsolateral prefrontal cortex additionally comes into play to encode links among items. Ragland’s preliminary data point to a role of the latter in the relational memory deficits that occur in schizophrenia; in contrast, item-specific memory and the ventrolateral prefrontal cortex remain more intact.
Further emphasis on the role of the prefrontal cortex came from Barch, who suggested that the region interacts with the hippocampus to control episodic memory. Many studies of episodic memory have used intentional encoding paradigms: for instance, they tell subjects to memorize words for a test without guiding them as to how to do so. This may work against subjects with schizophrenia who, compared to controls, tend to use less effective encoding strategies, such as careful reading of the material to be remembered.
To examine this notion, Barch and colleagues tested the memory of individuals with schizophrenia and healthy controls under incidental as well as intentional encoding paradigms. The former oriented subjects to process the meaning of words, a strategy that narrowed the recall gap between patients and controls.
According to Barch, deficits in episodic memory in schizophrenia may stem, at least in part, from decreased use of effective encoding strategies, which are mediated by the dorsolateral prefrontal cortex. Using magnetic resonance imaging, she found that subjects with schizophrenia show greater activation of this region than controls do when oriented to use such strategies. Since less activation correlated with better performance, semantic processing may occur less automatically in schizophrenia. Furthermore, when left unguided as to strategy use, subjects with the disease appeared less able than healthy subjects to muster additional support from the hippocampus to bind items in memory.
Next up, John Serences, of the University of California at San Diego, gave an overview of the neural mechanisms that underlie the attentional processes that keep us from constantly being overwhelmed by stimuli. Using event-related fMRI, he found that attention affected particular regions of the contralateral visual cortex that encode information about the stimulus. In addition, he examined whether attentional control signals arise from top-down, goal-directed processes or from bottom-up, stimulus-driven ones. He found that shifting the focus of attention activates the parts of the contralateral visual cortex that represent the relevant stimuli. Voluntary shifts in attention also increase activity in the dorsal parietal lobe. On the other hand, stimulus-driven attentional control—which occurs, for instance, when an unexpected object commands attention—causes changes in the extrastriate visual cortex, ventral parietal cortex, and ventral frontal cortex.
Serences sees a top-down, dorsal network that contributes to voluntary, or goal-directed shifts of attention, and a ventral, stimulus-driven network that responds to objects that appear suddenly, seem out of place, or are potentially relevant. These top-down and bottom-up networks work together to keep us attending to the task at hand, but not so narrowly that we miss the grizzly bear in the room.
Attention has “been used, abused, and so confused” in the schizophrenia literature, according to Jim Gold, of the Maryland Psychiatric Research Center, in Catonsville, Maryland, who borrowed from musician Van Morrison to make his point. Clinical tests do not measure pure attention; they may also invoke other processes, such as executive control and working memory, which are also impaired in schizophrenia. Executive control helps us choose which rules govern attention at any given time; working memory stores both the rules and the information that is perceived. Although patients can use attention to control perception, they perform much worse than healthy controls when competition for attention increases and the situation requires top-down attentional control. Rather than a specific problem with control, patients’ performance on the A-X version of the Continuous Performance Test suggests that a deficit in working memory capacity, rather than impaired executive control, underlies their poor performance.
In what he called a “data light” talk, Angus MacDonald, from the University of Minnesota in Minneapolis, stressed theory instead—specifically, the implications of cognitive deficits for leading biological theories of schizophrenia. He wondered whether cognitive neuroscience can help researchers test these theories. For instance, the dopamine theory suggests that dysregulation of the neurotransmitter could cause subjects to assign inappropriate salience to neutral stimuli. Glutamatergic theories attribute reduced neuroplasticity in patients to abnormalities involving NMDA and AMPA receptors, which lead to abnormal connectivity between certain parts of the brain. Theories that emphasize GABA dysregulation also have implications for attention; for instance, healthy controls under the influence of ketamine, an NMDA receptor blocker, perform similarly to patients on an attentional control task. MacDonald challenged cognitive scientists to test these theories by using them to derive competing hypotheses about cognition.
Translating basic science into interventions for humans
The next two talks explored the challenge of translating basic cognitive neuroscience findings into treatments. Dwight Dickinson, VA Capitol Health Care Network, Baltimore, Maryland, noted that cognitive neuroscience suggests that drugs that act on the acetylcholine system might improve cognition in schizophrenia, but clinical trial results using traditional neuropsychological measures have proven disappointing. This might indicate a need for more biologically based measures that meet CNTRICS (Cognitive Neuroscience Treatment Research to Improve Cognition in Schizophrenia) criteria, he said. After all, neuropsychological tasks, despite their advantages, tend to gauge more than one component, and their connection to underlying biological processes seems unclear.
However, Dickinson argued, the CNTRICS criteria focus on component cognitive processes that may not reflect reality; these processes do not work alone but in concert with other cognitive processes, the environment, and biology. As in genetics, where research is revealing combinations of genes that interact with other genes, the environment, and subject characteristics to influence schizophrenia, it seems likely that component cognitive processes might interact with other processes to produce behavior. Furthermore, performance on different cognitive dimensions correlates more highly in subjects with schizophrenia than in controls, perhaps due to compensatory mechanisms. Neuropsychological tasks may reflect this integrated reality better than CNTRICS-type measures do.
The session ended with a talk by Richard Keefe, of Duke University in Durham, North Carolina, who spoke about how to measure cognition in clinical trials, especially those conducted at multiple sites. He advised researchers to select sites that not only have appropriate space and resources for testing, but also experience with tasks similar to the ones that will be used. Keefe noted that placebo effects in clinical trials have been growing because trial sites lack competence in administering psychiatric rating scales. He said this has not compromised existing cognitive studies, but the potential concerns him.
Keefe also urged researchers to adopt a process for training and certifying those who administer and score cognitive tests. To earn certification, testers should have to demonstrate their competence face-to-face. In addition, delays in launching enrollment should prompt them to undergo a refresher course and recertification. To avoid data catastrophes, researchers should review the quality of the data as they come in and provide regular opportunities for testers to discuss any problems that arise.—Victoria L. Wilcox.