The Neural Systems Lab
       Of the CCNY Neural Engineering Group
Research in the Kelly Lab is focused on measuring and characterizing perceptual and cognitive brain signals that relate to behavior. We are particularly interested in linking non-invasively recorded electrical brain signals (EEG) in humans to specific neural computations involved in perception, attention and decision making. Our focus is on paradigm development and basic research findings, but we continually deploy our paradigms and dependent measures to studies of neurological and psychiatric disorders through several active clinical collaborations.
Our current research questions and projects include:
The "details" of attention: when you know something about a sensory event before it happens, you're better at parsing that event to get the information you need, for the situation you're in. This ability is known as selective attention. We know that the underlying brain mechanisms involve alteration of activity in the brain's sensory hardware. But the details of this alteration remain a mystery. For example, under what special conditions will attention cause a change in neural activity at the very moment sensory information hits the cortex (see Kelly et al 2008)? We are currently resolving these details through quantitative characterization of behavior and careful manipulation of task demands, drawing on principles of sensory physiology and computational models of perceptual decision making.
Neural mechanisms for perceptual decisions: When interacting with our sensory environment, we need to be sure about what we're looking at or hearing in order to decide how to act. This often requires gathering sensory evidence over time until a criterion level of certainty is reached. In collaboration with Redmond O'Connell of Trinity College Dublin, we have developed paradigms that allow us to isolate neural signals responsible for 1) encoding the relevant sensory quantity that is being decided upon (e.g. color or brightness), 2) integrating that information into a decision variable that traces the likelihood that the emerging decision is correct, and 3) preparing to act upon that decision. These are the three major stages of brain processing intervening between sensation and action, and our ability to provide a window onto each in isolation has considerable implications for understanding brain dysfunction. We are collaborating with Felice Ghilardhi and Clara Moisello of Sophie Davis School of Biomedical Education, CCNY to apply this approach in Parkinson's Disease (PD), Parkinson's with dementia (PDD) and dementia with Lewy Bodies (DLB); and we are collaborating with Michael Milham of the Child Mind Institute to apply the approach in neurodevelopmental disorders.
Neural mechanisms for value-based decisions: We make choices according to the relative value of competing options. How does the brain represent the spatial positions of objects of different value? How does it represent how much more valuable one object is than another? We are currently characterizing electrical brain signals that provide such representations, and establishing their dynamics and association with actions used to acquire the valuable objects.
Visual signals for contrast perception: The amplitude of visual activity driven by inputs of varying intensity and location can be measured efficiently by flickering visual stimuli in particular ways. These stimulation techniques are invaluable for examining fundamental visual processing mechanisms in humans and assessing deficits in the clinic. We are developing ways to enhance these measures by exploiting basic principles of sensory physiology and employing elementary signal processing. Our technique provides more robust indices of contrast perception, allowing us to examine the basic issue of how signals in visual neurons relate to what an observer reports seeing, and the influence of inhibitory neural interactions on visual perception. In collaboration with Fred Lado and Stephan Bickel of Montefiore Medical Center, we are applying this approach to the study of epilepsy, a condition related to neuronal over-excitation.
The origin of the brain's "system failures": Even when doing simple tasks, behavioral performance fluctuates and errors occur. Sometimes the brain's sensory analyzers are just not good enough. Other times, we just drift off the task through mind-wandering or simply because we momentarily loosen our focus. What's going on in the brain when this happens, and to what degree does it depend on the nature of the task at hand? In certain continuous monitoring situations, neural signatures of an impending error are evident in global brain signals as much as 20 seconds before the error occurs (see O' Connell et al 2009). How specific are these error-predicting signals to the parameters of our continuous monitoring task - do they generalize to the corresponding situations in real life, e.g. when driving a car?