Adaptive allocation of attentional gain

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Human sensory systems are continuously bombarded with far more input than they can process. As a result, attentional mechanisms have evolved so that available capacity is dedicated to encoding only the most salient and behaviorally relevant stimuli in the environment. In turn, the most important stimuli dominate perceptual awareness and have privileged access to memory stores and to the neural mechanisms that control decisions about how to best interact with external objects. In this proposal, we will use the visual system as a model to understand the basic brain-behavior processes involved in selective attention influence perception, working memory, and the computation of sensorimotor decisions. In addition, we will develop and employ new methods that use fMRI to non-invasively study attentional modulations and the information encoding capacity of sensory systems, in line with the strategic aim of the NIMH to develop novel tools and methodologies for understanding how populations of neural cells work together within and between brain regions. Traditional accounts hold that attention operates to magnify the neural response evoked by important stimuli, which makes a stimulus easier to perceive. This general framework is intuitive, and has been successfully guiding empirical studies for more than three decades. However, recent theoretical work suggests that attention should not simply increase the gain of neurons tuned to a relevant stimulus. Instead, attention should modulate the activity of sensory neurons in a more dynamic manner in order to maximize the probability that a specific perceptual task will be successfully completed. Often times, this counterintuitively requires enhancing the activity of neurons that are most responsive to stimuli that are not physically present in the visual field, because these neurons carry more information about very difficult discriminations between similar items (e.g. when a radiologist searches for a cancerous mass in a low-quality x-ray image). Recent empirical studies support this general framework, and further raise the intriguing possibility that individual differences in the optimality of attention can predict overall performance on difficult discriminations as well as the ability to improve on difficult discriminations with practice (learning). Here, we will critically evaluate this new theoretical perspective, and we will also explore how differences in attention across individuals can influence the precision of short-term memory and the efficiency of simple decision making processes. Collectively, our goal is to provide insights into the operation of attentional mechanisms so that we can more precisely characterize how the system should ideally operate. In turn, this should dramatically improve our ability to isolate specific aspects of attentional processing that can sometimes go awry, thereby enabling more targeted diagnoses and interventions in clinical settings.