Kenway Louie

The decisional reference point, the central mechanism underlying behavioral economics, conditions our evaluations of all reinforcers. Whether a given event is experienced as positive or negative depends on this reference point. A pathological elevation of the reference point would lead a person to experience once pleasurable activities as negative reinforcers. Thus, it has been hypothesized that a reference point pathology may play a critical role in the symptomology of major depressive disorder (MDD). Here, we test the hypothesis that the reference point is pathologically elevated and dynamically inflexible in patients suffering with MDD compared to healthy controls. We find that depression is associated with a significant elevation of the reference point, and the magnitude of this elevation correlates with disease severity. The ability of patients with MDD to dynamically adjust their reference point to the environment is also dysfunctional. Our findings link the previously demonstrated treatment of depression by deep brain stimulation to modulation of the reference point in the anterior cingulate cortex and identify pathology in the dynamics of reference point setting as a potential cognitive mechanism in the disorder. Finally, these results reveal that a three-minute video game-like task measuring the reference point strongly correlates with depression severity. After further testing for clinical validity, this rapid assay may serve as a potential tool for assessing and monitoring depression.

The decisional reference point serves as a hidden benchmark for evaluating options in decision-making. Despite extensive behavioral evidence for the existence of the reference point, its neural instantiation remains unclear. To identify reference point encoding at both the single-neuron and population levels, we analyzed neural activity from macaques performing a wealth accumulation task that dissociated objective reward values from the reference point. Across six frontal regions, we found that reference-related signals were broadly distributed at the single-neuron level. However, at the population level, only the ventral bank of the anterior cingulate cortex (vbACC) encoded the reference point significantly. In contrast, the dorsal bank of ACC and dorsolateral prefrontal cortex showed population-level encoding of reference-dependent subjective value signals. The temporal sequence of these signals and their known anatomical connectivity hints at a dedicated neural circuit for reference dependence, with the vbACC potentially serving as a global source of reference point signal modulating activity in downstream value-encoding regions.

The Divisive Normalization (DN) function has been described as a “canonical neural computation” in the brain that achieves efficient representations of sensory and choice stimuli. Recent theoretical work indicates that it efficiently encodes a specific class of Pareto-distributed stimuli. Does the brain shift to different encoding functions in other types of environments, or is there evidence for DN encoding in other types of environments? In this paper, using a within-subject choice experiment, we show evidence of the latter. Our subjects made decisions in two distinct choice environments with choice sets either drawn from a Pareto distribution or from a uniform distribution. Our results indicate that subjects’ choices are better described by a divisive coding strategy in both environments. Moreover, subjects appeared to calibrate a DN function to match, as closely as possible, the actual statistical properties of each environment. These results suggest that the nervous system may be constrained to use divisive representations under all conditions. Significance Statement How does the frequency with which we encounter different kinds of decision problems affect how the brain represents those problems? Recent empirical findings suggest that we adapt our internal representations to match the environments in which we are making choices. Theoretical work has shown that one form of internal representation, called divisive normalization, provides an optimal adaptation when making choices in a specific class of environments. Using a stylized experimental design, subjects faced two distinct choice environments, each characterized by different statistical properties. Our findings show humans appear to use the same mechanism in both environments, suggesting that a divisive representation may be a fixed feature of human cognition.

Noise is a fundamental problem for information processing in neural systems. In decision-making, noise is assumed to have a primary role in errors and stochastic choice behavior. However, little is known about how noise arising from different sources contributes to value coding and choice behaviors, especially when it interacts with neural computation. Here we examine how noise arising early versus late in the choice process differentially impacts context-dependent choice behavior. We found in model simulations that early and late noise predict opposing context effects: under early noise, contextual information enhances choice accuracy; while under late noise, context degrades choice accuracy. Furthermore, we verified these opposing predictions in experimental human choice behavior. Manipulating early and late noise – by inducing uncertainty in option values and controlling time pressure – produced dissociable positive and negative context effects. These findings reconcile controversial experimental findings in the literature reporting either context-driven impairments or improvements in choice performance, suggesting a unified mechanism for context-dependent choice. More broadly, these findings highlight how different sources of noise can interact with neural computations to differentially modulate behavior. Significance The current study addresses the role of noise origin in decision-making, reconciling controversies around how decision-making is impacted by context. We demonstrate that different types of noise – either arising early during evaluation or late during option comparison - leads to distinct results: with early noise, context enhances choice accuracy, while with late noise, context impairs it. Understanding these dynamics offers potential strategies for improving decision-making in noisy environments and refining existing neural computation models. Overall, our findings advance our understanding of how neural systems handle noise in essential cognitive tasks, suggest a beneficial role for contextual modulation under certain conditions, and highlight the profound implications of noise structure in decision-making.

Inhibition is crucial for brain function, regulating network activity by balancing excitation and implementing gain control. Recent evidence suggests that beyond simply inhibiting excitatory activity, inhibitory neurons can also shape circuit function through disinhibition. While disinhibitory circuit motifs have been implicated in cognitive processes, including learning, attentional selection, and input gating, the role of disinhibition is largely unexplored in the study of decision-making. Here, we show that disinhibition provides a simple circuit motif for fast, dynamic control of network state and function. This dynamic control allows a disinhibition-based decision model to reproduce both value normalization and winner-take-all dynamics, the two central features of neurobiological decision-making captured in separate existing models with distinct circuit motifs. In addition, the disinhibition model exhibits flexible attractor dynamics consistent with different forms of persistent activity seen in working memory. Fitting the model to empirical data shows it captures well both the neurophysiological dynamics of value coding and psychometric choice behavior. Furthermore, the biological basis of disinhibition provides a simple mechanism for flexible top-down control of the network states, enabling the circuit to capture diverse task-dependent neural dynamics. These results suggest a biologically plausible unifying mechanism for decision-making and emphasize the importance of local disinhibition in neural processing.

Learning is widely modeled in psychology, neuroscience, and computer science by prediction error-guided reinforcement learning (RL) algorithms. While standard RL assumes linear reward functions, reward-related neural activity is a saturating, nonlinear function of reward; however, the computational and behavioral implications of nonlinear RL are unknown. Here, we show that nonlinear RL incorporating the canonical divisive normalization computation introduces an intrinsic and tunable asymmetry in prediction error coding. At the behavioral level, this asymmetry explains empirical variability in risk preferences typically attributed to asymmetric learning rates. At the neural level, diversity in asymmetries provides a computational mechanism for recently proposed theories of distributional RL, allowing the brain to learn the full probability distribution of future rewards. This behavioral and computational flexibility argues for an incorporation of biologically valid value functions in computational models of learning and decision-making.

Learning is widely modeled in psychology, neuroscience, and computer science by prediction error-guided reinforcement learning (RL) algorithms. While standard RL assumes linear reward functions, reward-related neural activity is a saturating, nonlinear function of reward; however, the computational and behavioral implications of nonlinear RL are unknown. Here, we show that nonlinear RL incorporating the canonical divisive normalization computation introduces an intrinsic and tunable asymmetry in prediction error coding. At the behavioral level, this asymmetry explains empirical variability in risk preferences typically attributed to asymmetric learning rates. At the neural level, diversity in asymmetries provides a computational mechanism for recently proposed theories of distributional RL, allowing the brain to learn the full probability distribution of future rewards. This behavioral and computational flexibility argues for an incorporation of biologically valid value functions in computational models of learning and decision-making.

We are constantly faced with decisions between alternatives defined by multiple attributes, necessitating an evaluation and integration of different information sources. Time-varying signals in multiple brain areas are implicated in decision-making; but we lack a rigorous biophysical description of how basic circuit properties, such as excitatory-inhibitory (E/I) tone and cascading nonlinearities, shape attribute processing and choice behavior. Furthermore, how such properties govern choice performance under varying levels of environmental uncertainty is unknown. We investigated two-attribute, two-alternative decision-making in a dynamical, cascading nonlinear neural network with three layers: an input layer encoding choice alternative attribute values; an intermediate layer of modules processing separate attributes; and a final layer producing the decision. Depending on intermediate layer E/I tone, the network displays distinct regimes characterized by linear (I), convex (II) or concave (III) choice indifference curves. In regimes I and II, each option’s attribute information is additively integrated. In regime III, time-varying nonlinear operations amplify the separation between offer distributions by selectively attending to the attribute with the larger differences in input values. At low environmental uncertainty, a linear combination most consistently selects higher valued alternatives. However, at high environmental uncertainty, regime III is more likely than a linear operation to select alternatives with higher value. Furthermore, there are conditions where readout from the intermediate layer could be experimentally indistinguishable from the final layer. Finally, these principles are used to examine multi-attribute decisions in systems with reduced inhibitory tone, leading to predictions of different choice patterns and overall performance between those with restrictions on inhibitory tone and neurotypicals.

Consumer valuations are shaped by choice sets, exemplified by patterns of substitution between alternatives as choice sets are varied. Building on recent neuroeconomic evidence that valuations are transformed during the choice process, we incorporate the canonical divisive normalization computation into a discrete choice model and characterize how choice behaviour depends on both size and composition of the choice set. We then examine evidence for such behaviour from two choice experiments that vary the size and composition of the choice set. We find that divisive normalization more accurately captures observed behaviour than alternative models, including an example range normalization model. These results are robust across experimental paradigms. Finally, we demonstrate that Divisive Normalization implements an efficient means for the brain to represent valuations given neurobiological constraints, yielding the fewest choice errors possible given those constraints. This paper was accepted by Elke Weber, judgment and decision making.