W Todd Maddox - US grants

DESCRIPTION (Adapted from applicant's abstract): Historically, research that examines cognitive deficits in neurological patient populations has remained separate from research that uses mathematical models to examine "normal cognition. However, these two areas have begun to merge (e.g., Farah, 1990; Maddox, et.al., 1995). Model-based analyses are often superior to traditional specific cognitive deficit, and (c) determine the magnitude of these deficits. Maddox, el. al. (1995) utilized modeling techniques to examine PD patients' ability (a) to integrate information from two aspects of a stimulus (termed dimensional integration, DI), and (b) to attend selectively to one aspect of a stimulus while ignoring the other (termed selective attention; SA). We found that PD patients showed no DI deficits, but a large proportion of PD patients did show SA deficits. This proposal outlines two extensions of this research. Project 1 will examine the differential effects of stimulus separability and integrality on PD patients' SA performance. Normal subjects' SA performance is strongly influenced by stimulus separability and integrality; SA is easy with separable dimension, but is hard with integral dimensions (Maddox, 1992). Maddox et al. (1995) utilized integral dimensions and found SA deficits in PD patients. In Project 1, subjects will participate in two experiments; one that uses integral dimensions, and the other that uses separable dimensions. Model-based analyses will be applied to isolate and quantify the effects of stimulus integrality on PD patients' SA performance. Project 2 will examine the differential effects of an internal vs. external criterion on PD patients' DI performance. Maddox et al. (1995) found that PD patient were not impaired in their DI ability when no internal referent was required. Some have argued that PD patients had more impaired on tasks that place greater demands on the internal control of attention (e.g., Brown & Marsden, 1988). Subjects will participate in two experiments like those used in Project 1, however one will require an external criterion, and the other will not model-based analyses will again be used to isolate and quantify the effects of an internal vs. external criterion on PD patients' DI performance.

DESCRIPTION (Applicant's abstract): The long-term objective of the proposed research is to better understand the nature of the perceptual and cognitive processes involved when a person identifies or categorizes objects. Two lines of research are outlined. The first examines the optimality of human categorization performance when base-rates and payoffs are manipulated simultaneously. Experiments are outlined that examine the effect of (a) stimulus properties, (b) category discriminability, and (c) the costs and benefits associated with various categorization responses on base-rate and payoff sensitivity. All experiment will use the perceptual categorization task (e.g., Ashby & Cott, 1988; Maddox, 1995) in which two normally distributed categories are specified and a large number of category exemplars are sampled from each category distribution. The use of normally distributed categories allows the optimal decision rule to be derived. The approach is to apply a series of quantitative models to the trial-by-trial learning data and to measures of asymptotic performance. Each model will embody specific hypotheses about the optimality, or potential sub-optimality, of responding. Standard categorization procedures as well as hypothetical medical diagnosis procedures will be utilized. These studies will inform many real-world categorization problems, such as medical diagnosis, where base-rates and payoffs often differ and vary simultaneously. The second line of research examines perceptual and decisional attention processes in identification and categorization. Experiments are outlined that examine the effects of (a) attention cue and decision rule manipulations, as well as (b) stimulus property and response deadline manipulations on perceptual and decisional attention processes. To better understand the nature of these attention systems, attempts will be made to account simultaneously for accuracy and response time (RT) data within a single theoretical framework (e.g., Maddox & Ashby, 1996). To ensure stable estimates of the accuracy and RT data, each task will utilize a small number of unique stimuli (15-30) with 200-300 presentations of each stimulus in each condition. The approach is to apply a series of quantitative models that each embody specific hypotheses about the nature of, and interaction between, perceptual and decisional attention systems, and to compare these models with models derived from theories that do not distinguish between perceptual and decisional forms of attention (e.g., many exemplar-based models; Nosofsky, 1986). These studies are important because they will provide useful information about attention processes that are fundamental to nearly all types of human behavior.

DESCRIPTION (provided by applicant): The long-term objective of the proposed research is to identify and quantify the perceptual and cognitive processes that are involved when an observer is presented with a categorization problem in which the prior probabilities (or base-rates) of the categories, and the costs and benefits (or payoffs) associated with categorization decisions are manipulated. With funding from NIH Research Grant # S R01 MH59196 my students and I made significant progress toward understanding the processes involved in decision criterion learning when base-rates and payoffs are manipulated, and toward understanding the complex interplay between several factors that influence base-rate/payoff learning. This work answered many questions, but also suggested many new lines of research. The purpose of this proposal is to expand our previous work in several new directions. The approach taken in the proposed research is to isolate and quantify the influence of several variables on decision criterion learning by comparing human performance with that of the optimal classifier--a hypothetical device that maximizes long-run reward. The aim is to test quantitative models of trial-by-trial and asymptotic performance by developing an "optimal" and several "sub-optimal" models, which instantiate important theoretical constraint. Four lines of research are proposed. Project 1 examines the effects of category distribution manipulations on base-rate and payoff learning. Theoretical work suggests that category discriminability, d', and category variance manipulations have a large effect on the rate of change in reward (or steepness) of the objective reward function which relates objective reward to the location of the decision criterion. If observers are sensitive to differences in steepness (called the flat-maxima hypothesis) then this should affect the speed and asymptote of learning. Project 2 examines the effects of payoff matrix manipulations on decision criterion learning. Theoretical work from our lab suggests that payoff matrix multiplication affect steepness, whereas matrix addition does not. Project 3 examines different types of feedback that might improve decision criterion learning. Especially promising is feedback based on the optimal classifier. Project 4 extends the studies in Project I - 3 to an explicit decision criterion task where observers adjust an observable decision criterion on each trial. These data are useful for testing learning models.

DESCRIPTION (provided by applicant): Psychology makes a distinction between motivation-processes that drive an individual to act-and cognition-processes by which information is processed. Despite their separation within Psychology, research on motivation and cognition need to be brought together because there is no cognition in the absence of motivational influences. Furthermore, cognitive neuroscience and clinical neuropsychology suggest that the brain areas responsible for motivational influences are not anatomically or functionally separable from those responsible for information processing. Our proposed work reunites research on motivation and cognition. This goal is crucial for understanding of normal functioning and for our ability to understand and treat cognitive deficits in clinical patients. Our motivational framework (derived from regulatory focus theory) assumes that people's motivational states can be focused on potential gains (a promotion focus) or on potential losses (a prevention focus). Our emphasis in on motivational influences on classification learning. Classification learning provides an ideal testbed for our studies because (a) much is known about the neurobiological systems and cognitive processes involved, (b) these neurobiological systems overlap extensively with those implicated in patients with clinical disorders, and (c) the Pis have over 25 years of combined experience in this field. The specific aims are to examine the effects of regulatory focus on explicit hypothesis-testing learning and implicit similarity-based learning. We also introduce social focus into the regulatory focus-learning framework. Social motivational factors are likely critical to an understanding of many neuropsychological disorders (e.g., anxiety and depression). The public health implications of this work are many. First, without understanding normal functioning, we cannot determine whether clinical patients (e.g., those with anxiety, depression, schizophrenia, etc) perform poorly because their disorder leads to cognitive impairments, or because it leads to a motivational mismatch. Second, a more detailed understanding of the motivation-learning interface will lead to improved neuropsychological testing measures and rehabilitation training strategies. Little is known about the motivational factors in clinicial disorders and about the motivation-cognition interface. This proposal reunites research on motivation and cognition to better understanding their effects on functioning in clinical populations.

DESCRIPTION (provided by applicant): Deficiencies in reward and punishment processing are theoretical cornerstones of alcohol and substance dependence, addiction, and other psychopathology. However, most research ignores the fact that contemporary cognitive theory emphasizes two processing modes: a reflective mode where processing is under conscious control and predominantly frontally-mediated, and a reflexive mode that is not under conscious control and is predominantly striatally-mediated. In addition, a detailed understanding of the genetic underpinnings of dual processing modes of reward and punishment is critical to improving our theories of addiction and psychopathology and to translational work focused on developing interventions. Dopamine and serotonin genes are hypothesized to affect reflexive and reflective reward and punishment processing and are therefore the focus of this proposal. The overall goal of this project is to test specific hypotheses regarding dopaminergic and serotonergic genetic variation on reflexive and reflective reward and punishment processing. We use classification learning tasks for which the optimal mode of processing (reflective or reflexive) can be defined rigorously and for which the research team has over 20 years of experience. The proposed studies will also complement a single nucleotide polymorphism approach with a haplotype strategy, which will determine whether additional variants in these dopaminergic and serotonergic genes also influence reward and punishment processing. We will also account for population stratification by testing and statistically controlling for occult population substructure. Aim 1 examines associations between genetic variation in dopaminergic and serotonergic systems with reward and punishment processing when optimal performance is mediated by the reflexive system or by the reflective system. Aim 2 examines the effects of reflexive system genetic variation on reflective-optimal task performance, and reflective system genetic variation on reflexive-optimal task performance. Aim 3 examines the influence of stress on reflexive reward and punishment processing. The proposed studies are the first to attempt to characterize the interactive effects of serotonin, dopamine and stress on cognitive processing of rewards and punishment using contemporary cognitive frameworks. This integrative, interdisciplinary research approach will provide the critical foundation needed for future translational work that examines how these processes go awry in clinical disorders.

DESCRIPTION (provided by applicant): Age-related deficits in decision making are well documented and these decisions often have huge social implications for individuals and their family members, leading to significant pressure to make the best decisions. Much of the previous research on aging focused on history-independent decision-making tasks for which current rewards are independent of previous decisions. This research ignores situations where the current rewards available from each option are influenced by previous decisions. History-dependent decisions are ubiquitous and they are often accompanied by various forms of pressure. Recent work in our labs suggests that older adults perform better in history-dependent situations while younger adults perform better in history- independent situations. However, the precise neural and computational mechanisms associated with these age-based differences remain unclear. Additionally, virtually no research has examined how older adults respond to pressure despite its prevalence. Recent work suggests that normal aging is associated with declines in the neuromodulation of the frontostriatal limbic network associated with decision-making. Other work provides evidence for compensatory over- activation in brain regions, particularly lateral frontal brain regions, for older relative to younger adults in a variety of cognitive tasks. This over-activation is seen as compensatory for neural declines associated with aging and is known as the compensation-related utilization of neural circuits hypothesis (CRUNCH). We hypothesize that compensatory over-activation can account for the age-related advantage in history-dependent decision-making. In addition, compensation related frontal activity in older adults may follow an inverted U- shape as cognitive demand increases. With increased cognitive demand, under-activation in older adults, relative to younger adults might result when the crunch point is reached. Increased pressure in decision- making situations may force older adults to hit such a crunch point. The goal of this proposal is to examine the effects of aging on history-dependent and history-independent decision-making, and to systematically test predictions of the CRUNCH hypothesis as it applies to decision- making under pressure. Our research team is highly qualified to achieve these aims given our expertise in brain imaging, computational modeling, and behavioral studies of normal aging and pressure. We will apply models that assume qualitatively different strategies to the data and prediction errors from these models will be used as regressors with neural activity. We predict that older adults will show greater activation in DLPFC and LOFC, compared to younger adults under no pressure conditions but that older adults will show under- activation in these same regions under pressure conditions. Aims 1 and 2 examine the neurobiological underpinnings of age-related changes in history-dependent and history-independent decision-making. Aim 3 extends Aims 1 and 2 by examining decision-making under pressure.