Barbara J. Knowlton - US grants

memory disorders; performance; association learning; verbal conditioning; discrimination learning; psycholinguistics; brain injury; human subject;

Lay Abstract: Mnemonic Functions of the Basal Ganglia

A great deal of research points to the idea that there are different forms of memory which depend on different brain systems. Explicit memory, which is memory for facts and events that you are consciously aware of appears to depend on brain circuitry including medial temporal lobes structures. Implicit memory, on the other hand, refers to information learned without awareness, which does not seem to depend on this brain system. In the proposed project, we plan to study habit learning, which is a form of Implicit learning. Habit learning refers to the gradual learning of associations across trials. Some research has suggested that the basal ganglia play a role in this type of learning. However, habit learning is not well understood in terms of the precise neural circuitry underlying it or its behavioral characteristics. By using animal models and studies with patients with basal ganglia disorders I plan to address both of these questions. In one habit learning task, we train rats to find food in parts of a maze that are signaled by a visual cue. We plan to make lesions in the caudate nucleus (part of the basal ganglia) at different times after training to see whether memories are stored in this structure or if it is only involved early in training with memories eventually stored elsewhere.
We will also investigate the role of the nucleus accumbens in habit learning, a structure with significant interconnections with the basal ganglia that has been implicated in reinforcement. The work on humans will focus on patients with Parkinson's disease, which results in basal ganglia dysfunction. We will use a human habit learning task that I have developed in which different cues are probabilistically associated with different outcomes. Because the associations are probabilistic, subjects learn them implicitly and feel as if they are guessing. Patients with Parkinson's disease are impaired at this task, and I will test whether surgical treatments used to treat this disease have any effect on performance. I will also develop additional tasks to tap into this system, and examine the behavioral characteristics of these tasks, such as their dependence on attention. I will compare patients with Parkinson's disease and Alzheimer's disease on habit learning and explicit memory tasks. Because Alzheimer's disease mostly spares the basal ganglia, I will examine if habit learning can proceed normally in those patients. I will also look at basal ganglia activity using functional MRI during performance of implicit learning tasks to help define the characteristics of basal ganglia dependent learning. The proposed project also has a large educational component, in that I am continuing to develop a cognitive neuroscience major in our Psychology department to give students integrated training in the study of the nervous system and in cognitive science.
Understanding the functions of basal ganglia has practical significance because this system is affected in a number of human disease states including Huntington's disease and Parkinson's disease. Although these diseases are generally thought of as involving motor function, they also involve cognitive deficits and mood changes. Implicit learning problems may have an impact on these deficits, and the depression associated with these diseases may in part stem from difficulties in using implicitly learned social information. Learning how the basal ganglia circuitry participates in learning will help us to better treat patients with these diseases.

Three studies will use functional magnetic resonance imaging (fMRI) to assess the neuroanatomical circuitry involved in three social cognitive domains. In one study, the neuroanatomical of schematic social information will be examined. In a second study, the neuroanatomical differences associated with high-level abstract understanding versus low-level concrete understanding of events derived from action identification theory will be investigated. The third study will focus on the neuroanatomical correlates of outcome framing in terms of gains versus losses. This is a small grant for exploratory research aimed at establishing the utility of a cognitive neuroscience analysis of social cognitive phenomena.

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).

A major advance in the study of the brain has been the discovery of specific regions that play a critical role in memory. These regions include the medial temporal lobe, which includes several different interconnected structures. With support from the National Science Foundation, Dr. Barbara Knowlton and colleagues at UCLA will use functional magnetic resonance imaging to discover the different roles of these regions in human memory, and how they work together to support the rich, detailed memories of our experiences that define us as people. Much of the work is motivated by theories of human memory that state that one medial temporal lobe region, the CA3 region of the hippocampus, is very important for the creation of distinct memories for events, and allows us to keep memories for similar events from blending together. Dr. Knowlton and her research team will test the idea that activity in the CA3 during learning is directly related to how robust and long-lasting memories will be. The experiments will also compare brain activity when people learn information by trying to think about what is unique about each item to brain activity when people learn information by concentrating on what is similar across items. The research hypothesis is that there will be more activation in the CA3 region in the first kind of learning because the brain will be forming distinctive memories. Additionally, the research will test the idea that the same pattern of activity that occurs in the medial temporal lobe when an item is learned will be repeated when that same item is later remembered.

These experiments will provide important new information about how neural circuits form human memories, and thus will be the source of insights into how memory can be improved. Furthermore, this proposal will train graduate students in high resolution and cortical unfolding techniques, which can be applied to other areas of cognitive neuroscience.

Skill learning is important in areas that enrich our lives; however, people typically do not practice skills in the optimal manner for learning. For example, people often practice one variation of a skill over and over again, whereas practicing different variations of the skill mixed together leads to much better retention. This project will examine how this intermixed practice may be beneficial to skill learning that can generalize to new similar skills, and will investigate how to best practice a skill to support such transfer. Participants in these studies will learn simple finger tapping skills, similar to what one learns while playing piano or typing. Functional magnetic resonance imaging (fMRI) will be used to examine brain areas that are more active when people are practicing in a way that leads to better generalization to similar skills. This information will be used to improve skill learning using transcranial direct current stimulation (tDCS), which involves applying low levels of current to brain regions from the surface of the scalp. People will learn motor skills while receiving this stimulation to see if it improves their ability to later learn a new, similar skill. The findings from these studies may reveal important ways to enhance skill learning by improving the way we practice and enhancing brain activity to make people better at skill learning. As tDCS is a new, inexpensive and essentially risk-free treatment, the results here could be applied quite broadly. Given the importance of effective skill training to industry, the military, and daily life, what is learned from these studies could have a broad benefit to society by making skill training more efficient. This research will also impact education by supporting a graduate student who will learn neuroimaging and brain stimulation methods. The project will support research experiences for undergraduates drawn from the diverse population at UCLA and will make links with the community through the researchers involvement with programs designed to inspire local high school and elementary students to consider careers in life science.

This project seeks to understand how to engage the brain regions during learning that lead to the ability to transfer to new related skills. Subjects will practice a set of simple motor sequences using the serial reaction time task (SRTT) and will be tested on their ability to efficiently perform new sequences. We hypothesize that practicing multiple related motor skills in an interleaved manner will enhance transfer and engage different brain regions compared to when practicing these skills in a blocked manner, with the latter leading to interference between skills. In this study, subjects will practice interleaved sequences in the SSRTT in an fMRI scanner, and regions will be identified in which BOLD signal during learning is correlated with subsequent transfer across subjects. In this way, the components of the motor learning system can be identified that are involved in the formation of a memory trace that can support transfer. Pilot work has identified the cerebellum as a likely target, and in the final proposed study, tDCS will be used to stimulate cerebellum during SSRTT practice and subsequent transfer to new sequences. If it can be understood how interleaved practice leads to the creation of a memory trace that leads to better transfer, it would provide key insights into neural representations of memory.

Early life stress (ELS) is strongly associated with a host of negative health consequences in adulthood, including addiction and obesity. However, there is a critical gap in our understanding of the mechanism underlying these relationships.. We propose to test an innovative set of hypotheses linking behavioral and neurocognitive mechanisms of habit vs. goal-directed learning with ELS. We further propose to evaluate the role of conditional fear cues as potentiators of habit learning, via Pavlovian-Instrumental Transfer in individuals who experienced ELS. This study will test a model by which reminders of past fearful experiences enhance habit learning in individuals already vulnerable to addiction. These studies apply the rigor of learning theory concepts to understand behavioral vulnerabilities resulting from ELS. We will link work in experimental animals on neural circuits of habit learning with human subject studies using functional connectivity methods and neuroimaging. We will test a young, healthy population to maximize the isolation of risk factors associated with ELS prior to the onset of disease. Participants will be grouped based on no major early life stressors, one or two early life stressors, and three or more early life stressors, to evaluate the dose relationship between stress and propensity for habit learning over goal- directed learning. We propose to measure habit learning using an instrumental task that we have adapted for use in the fMRI scanner. By assessing whether the learned response is sensitive to devaluation, we will test whether learning is habitual vs. goal- directed. We hypothesize that individuals who experienced ELS will show greater rates of habit learning, and the presence of a fear CS will accentuate this difference between the groups. We will use the same behavioral paradigm to assess neural mechanisms of habit responding using high-resolution fRMI that will enable us to assess BOLD signal activation in striatal subregions during instrumental learning. Based on work in experimental animals, we hypothesize that individuals who experienced ELS will show activation in more lateral regions of the striatum including the putamen, while in control participants, activation will be more medial, in the caudate nucleus. We will be able to achieve greater anatomical precision in our analyses by our use of our meta-analysis of striatal activation in habit learning tasks in the literature. We will also be able to assess functional connectivity with the amygdala and other regions, thus integrating our findings with work linking early-life stress to amygdala hypertrophy. An array of addiction-related behaviors (e.g., smoking, drug use, alcohol use ) will be tested as correlates of behavioral and neural indices of habit learning to validate the model.