Theresa M. Desrochers, Ph.D. - US grants

DESCRIPTION (provided by applicant): Many everyday tasks require sequences of multiple goals and subgoals. For example, the goal of making a sandwich may contain the subgoals of toasting the bread and spreading a condiment, which themselves consist of sequences of subgoals (e.g. getting a knife, opening the jar, etc.). Our ability to guide these actions flexiblyis known as cognitive control, and this is a primary function of the frontal cortex. Recent research has suggested that the frontal cortex is organized along the rostrocaudal axis to control cognitive tasks at different levels of abstraction (goals to sub-subgoals), or hierarchically. Most of this work, however, used non-sequential tasks; therefore, it has been only assumed that similar control mechanisms will support hierarchical control over sequences of tasks. This leaves at least three unanswered questions fundamental to our understanding of frontal lobe and cognitive control function, generally: 1) does the rostrocaudal organization of frontal cortex support hierarchical control of task sequences; 2) what are the dynamics of the control and monitoring throughout task sequences; and 3) how does the control of task sequences change with learning. These three questions will be addressed across our three aims by using functional magnetic resonance imaging (fMRI) to investigate the following hypotheses: 1) the hierarchical control of task sequences is supported by the rostrocaudal organization of frontal cortex; 2) monitoring will occur either in a sustained manner (throughout) or in a transient manner (at the beginning/end) for familiar task sequences; and 3) monitoring will transition from occurring in a sustained fashion over novel sequences to occurring transiently at sequence boundaries in familiar sequences with learning. The insight into the monitoring and control of task sequences that will be gained will be fundamental to the understanding of how we perform everyday tasks. This understanding is particularly crucial to the quality of life of those patients who have fronta lobe dysfunction or damage. Such patients often cannot live independently because they have trouble planning and monitoring the task sequences that are commonly required. PUBLIC HEALTH RELEVANCE: Patients with frontal cortical damage or dysfunction often are unable to live independently due to their inability to properly monitor and control the sequences of tasks that are common in everyday life. Although much is known about the mechanism by which the frontal cortex exerts this control over non-sequential tasks, comparatively little is known about the dynamics and structure of the brain areas that subserve sequences of tasks. The current project aims to directly investigate and broaden our knowledge of the neural systems that support the control of task sequences, which would provide the basis for progress towards frontal patients having a higher quality of life with better diagnoses and improved therapies.

PROJECT SUMMARY / ABSTRACT We perform sequences of tasks every day, such as making a cup of coffee. These `abstract' sequences can be distinguished from motor sequences along a variety of dimensions. In contrast to a specific order of muscle activations (e.g. playing piano) found in typical motor sequences, abstract sequences consist of sub- goals that might be achieved by any of a number of actions. Patients with frontal lobe damage have deficits in completing every day task sequences, even when they perform normally on non-sequential tests of executive function. Despite the importance of abstract sequences for understanding human behavior in health and for treating disease, relatively little is known about their neural mechanisms. Systematically answering this question is the defining goal of our research program. Our previous work using a sequential decision making task revealed a novel and necessary dynamic where activity in the rostrolateral prefrontal cortex (RLPFC) increased from the beginning to the end of each abstract sequence (?ramped?) using fMRI and TMS in humans. We hypothesized this ramp represented the resolution of accumulating positional uncertainty through the sequence. However, also inherent to all of these tasks requiring complex decision making is monitoring, and placing in order, multiple variables. The goal of this project is to test the prediction that RLPFC and its associated network support sequence monitoring modulated by uncertainty; and, simultaneously, develop an animal model of this human cognitive process that can be used for future cross-species hypothesis testing. The experiments in Aim 1 with utilize human fMRI in two separate but complimentary experiments to investigate how the dynamics in the rostral frontal previously found to be necessary for sequential task performance are modulated during sequence monitoring and uncertainty. Aim 2 will utilize nonhuman primate fMRI and the same behavioral paradigm as in humans to establish an animal model of abstract sequence monitoring and directly test functional homology with humans. This study will be the first comparison of performance of a sequential task of this kind across both species. In Aim 3 we will determine the necessity of the signals revealed in Aims 1 and 2, with direct manipulation of the circuits using TMS in humans and neurotransmitter agonists in monkeys. Together, the proposed experiments will compose a unique, cross-species investigation of the neural basis of abstract sequence performance. Such investigation is necessary to understand the complex yet ubiquitous sequences of tasks that make up daily life, and that patients with frontal lobe damage and Parkinson's Disease often struggle with. This understanding could contribute to novel treatments and therapies for such disorders.

PROJECT SUMMARY The ultimate goal of this research program is to determine the neural mechanisms of sequence monitoring. This knowledge can directly contribute to understanding new treatments for disorders where sequential behaviors are disrupted, such as Obsessive-Compulsive Disorder (OCD). Daily, we monitor sequences of visual information such as the series of bus or train stops when looking for the correct exit. Sequence monitoring is the active process of tracking the order of subsequent ?states? or steps. Monitoring is distinct from other well-studied sequence processes, such as explicit memorization, or potentially more automatic behaviors such as a series of motor outputs (e.g., playing the piano) or statistical sequence learning. However, the monitoring aspects of sequence processing remain largely unknown. Knowledge of higher-order similarities across sequences (e.g., AAAB, &&&@), abstractions, can aid sequence monitoring. For example, understanding the steps required in each turn in a game, or the repetitive pattern in a poem or song, can improve our awareness and processing. Abstract sequence monitoring includes viewing the sequences that possess abstract structure, active monitoring of this structure, and response. Here, we will determine the neural representation of abstract sequential structure using passive structured sequence viewing. We hypothesize that a key element of sequence structure, position, is encoded in the lateral prefrontal cortex (LPFC) with phasic bursts that in aggregate create population ramping. These neural dynamics can uniquely ?tag? each serial position in the sequence, supporting monitoring. This prediction is based on prior literature, and on our discovery that increasing (ramping) blood oxygen level-dependent (BOLD) activity in the rostral LPFC of humans is necessary for sequence execution and underlies sequence monitoring (Desrochers et al., 2015a, 2019). During awake fMRI, we observed parallel BOLD ramping in the LPFC of monkeys performing a structured sequence viewing task. We will use these data to specifically target electrophysiological recordings (Desrochers et al., 2015b; Feingold*, Desrochers* et al. 2012). This viewing task also provides robust BOLD responses to viewed sequence deviants and avoids motor and decision-making confounds while providing experimental flexibility. By using these activity patterns to guide neural recordings, we remove the need for anatomy-based assumptions about cross-species homology and can localize data acquisition with sub-region specificity. We will systematically test the hypotheses that sequence position is characterized by successive phasic increases in neural spiking in LPFC (Aim 1), and that neural activity related to sequence position is modulated by the passage of time and reward expectation (Aim 2). Sequence monitoring is fundamental to many natural behaviors. These data are unique in being guided by, and directly relatable-to, fMRI mapping. While hypothesis- driven, any outcome of these recordings from fMRI-identified brain areas will advance our understanding of this crucial process.