Decoding the Nucleus Accumbens Core: A Deep Dive into Learning and Reward
The process of associative learning, by which the brain links sensory stimuli with specific motor behaviors and expected rewards, is fundamental to adaptation and survival. Within the intricate circuitry of the brain, the nucleus accumbens core (NAcc) emerges as a critical player in learning associations between sensory cues and profitable motor responses. This article delves into the NAcc, exploring its function, neural mechanisms, and implications for understanding behavior and certain disease states.
The Nucleus: The Cell's Control Center
Before diving into the specifics of the NAcc, it's important to understand the basics of a cell's structure. The nucleus is an organelle containing DNA, the genetic material of the cell. DNA is vital for cells to work and make more cells, and is key to life. Learning about the nucleus helps us figure out how to diagnose diseases and find cures. Jobs like pathologists and molecular biologists need to understand the nucleus to spot diseases, create treatments, and even make lab-grown meat.
The Nucleus Accumbens Core (NAcc): A Limbic-Motor Interface
Anatomical, neurochemical, and brain lesion data suggest that the NAcc plays a role in modulating the motivation to perform reward-oriented behaviors as a “limbic-motor interface”. The NAcc receives glutamatergic inputs from orbitofrontal/prefrontal cortex, basolateral amygdala, and hippocampus (areas involved with stimulus properties, preferences, and memories), while dopaminergic input is received from ventral tegmental area neurons. NAcc outputs include projections to the ventral pallidum, the dorsomedial thalamus (which projects back to the orbitofrontal cortex), pedunculopontine tegmentum, and a significant projection to dopaminergic areas of the midbrain.
Lesion and drug studies have demonstrated that disruption of the NAcc results in decreased goal-directed behavior, dysfunction of reward encoding and learning as well as reduction in locomotor and approach behaviors. Correspondingly, dysregulation of the NAcc has been implicated in a number of disease states including major depression, drug addiction, and Parkinsons disease. One potential explanation for the above findings, the “incentive salience” hypothesis, posits that dopamine signaling via the mesolimbic dopaminergic pathway (which partially includes the NAcc) regulates motivation by associating values with environmental stimuli that predict reward. Moreover, during classical conditioning, the repetitive pairing of an external stimulus (e.g., visual, auditory, tactile) with a reward prompts increased firing rates of NAcc phasically active neurons (PAN's) during stimulus presentation. In contrast, when rewards are omitted, following previously conditioned stimuli (extinction), firing rates attenuate during stimulus presentation.
Visual-Motor Associative Learning Task
Unlike the reflexive responses of classical conditioning, operant conditioning requires formation of associations between external stimuli and spontaneously generated, volitional behaviors that result in reward. Furthermore, the mechanisms that promote reinforcement of profitable associations and attenuation of unprofitable associations in operant conditioning remain poorly understood. Thus, the activity of NAcc neurons in non-human primates as they performed a visual-motor associative learning task was examined. The primates focused on a central point on the screen until an object appeared (Stimulus). After a variable delay, the fixation point disappeared (GoCue), at which point the monkey was required to make a saccade from the center of the screen to one of four targets (Movement). An auditory tone (Feedback/tone) and color change of the selected target indicated whether the animal made the correct or incorrect choice. The former was followed by juice administration (Reward).
Read also: Understanding PLCs
Each trial began with the presentation of a central fixation point (0.2° diameter) surrounded by four gray targets (1° diameter and 10° from the center). Animals were required to fixate within 2° of the fixation point for 500 ms. Then either a novel or familiar stimulus appeared for 500 ms at the center, with the fixation point still visible. After a variable delay of 500-1000 ms, the fixation point disappeared, at which point the monkey was required to make a saccade from the center of the screen to one of the four targets. Once the animal fixated on a target for 500 ms, an auditory tone and a color change of the selected target indicated whether the animal made the correct (high pitch) or incorrect (low pitch) choice. A correct choice was followed by a juice reward after an additional 500 ms delay. An incorrect choice was followed by no reward. If at any point the animal failed to meet these criteria, the trial was aborted, and no reward was given. During each learning block, two novel stimuli (randomly generated geometric objects) and two familiar stimuli (randomly selected from a group of well-trained familiar objects with established movement directions) were presented. Each visual stimulus was associated with a unique saccade direction.
The use of the familiar objects served two important functions. First, familiar trials provide an impetus for the animals to continue working during the initial phase of learning, when correct choices for novel objects occur at a low frequency. Once the animals performed 16 correct trials for each object, the novel stimuli were replaced by two new randomly generated novel stimuli. This process was repeated multiple times for each neuron recorded, such that numerous instances of visual-motor associative learning were recorded for each neuron. Familiar and novel object trials were pseudo-randomly interleaved (i.e., each objects was randomly presented before any were repeated) within each block. Animals were trained on the behavioral task until they learned a minimum of four learning blocks per learning session. During the study, animals successfully learned 64% (n = 558/878) of novel object associations (to a 99% confidence interval) during the visual-motor association task and learned 4.7 ± 0.3 (mean ± s.e.m.) novel objects per recording session. On average, the animals learned novel associations in 10.0 ± 0.3 trials (mean ± s.e.m.; counting preceding incorrect and correct trials). Behavioral performance of the task demonstrated that the animals' performance started near chance (25%) and reached approximately 80% after learning occurred. Among familiar objects presented, animals selected the correct target in 98% of trials. Moreover, reaction times during the task were correlated with behavioral performance (p < 0.001; linear regression).
Two Classes of NAcc Neurons and Their Roles in Learning
During learning, responsive neurons can be divided into at least two distinct classes. The first class of neurons (Class I) exhibited a progressive increase in activity that was then maintained after novel visual-motor associations were mastered. These learning-related increases in activity were observed at the go-cue, feedback/tone and reward epochs of the behavioral task, suggesting a role in exploiting learned rewarded behaviors. In contrast, the second class of neurons demonstrated a decrease in activity that occurred only during the reward periods of the task. Hence, these “Class II” neurons may be involved in encoding profitable associations via down regulation of neuronal activity.
A total of 132 neurons were recorded from the NAcc from two non-human primates (monkey 1, n = 86; monkey 2, n = 46) as the animals performed the visual-motor association task. Of the 132 neurons recorded during the task, 88 (67%) were determined to be task responsive, and were further analyzed. The remaining neurons (n = 44/132) were classified as non-responsive and excluded from subsequent analysis. The aggregate median baseline firing rates (at the start of the trial) for task responsive neurons were 7 spikes/second (4-16 spikes/second quartiles; Table 1). Baseline median firing rates between Class I [6.9 spikes/second (4-13)] and Class II [7.1 spikes/second (4-20)] neurons were not significantly different (Mann-Whitney; p = 0.5). In addition, the aggregate mean discharge rate (Table 1) of responsive neurons demonstrated a significant increase in activity during the go-cue, feedback/tone, and reward epochs of the task (Friedman analysis of variance; p < 0.001, Dunn's correction). In order to evaluate neuronal activity in relation to learning, the series of correct and incorrect responses for each novel object was analyzed using a state-space approach to establish the trial at which an animal reached the learning criterion for a particular novel visual stimulus. This analysis approach provides the trial number (criterion trial) at which the animal's choice was statistically greater than chance at a 99% confidence interval. Responsive neurons were pooled into two groups based upon their correlation with the learning curve during the reward period of the task.
Familiar object trials do not require learning. These visual cues and their associated movement directions were presented to the animals thousands of times during training and were extremely well learned by the time of the experiment. Firing rate modulation during familiar trials demonstrated different patterns of activity between the two groups of responsive neurons. The population of Class I neurons (39 of 88 responsive neurons, 44%), responded to the behavioral task by a consistent increase in firing rate, compared to baseline, during the go-cue, feedback/tone and reward periods of the task (Table 1, Friedman analysis of variance; p < 0.001, Dunn's correction). Analysis of activity during novel object trials also revealed significant differences between the two classes of responsive neurons.
Read also: Learning Resources Near You
The learning-related activity of Class I and II neurons can be appreciated as representative neurons during a single learning event. The raster plots of a Class I neuron demonstrate a significant increase in activity in trials during and after learning at the go-cue, feedback/tone, and reward epoch of the task. In contrast, a Class II neuron had consistent discharge rates in all epochs of the task except for the reward period, during which it exhibited a decrease in activity near the learning criterion and afterward.
As a population, Class I neurons demonstrated a significant gradual increase in firing rates during the stimulus, go-cue, feedback/tone, and reward periods of the task as learning occurred. Neuronal activity prior to learning (trials, −10 to −7) was significantly lower than activity for familiar objects trials (at comparable epochs; repeated measures Freidman Analysis with multiple comparisons correction; X2Go(3) = 10.87, X2Tone(3) = 14.81, X2Reward(3) = 38.15, *p < 0.05, **p < 0.01, and ***p < 0.001). However, after learning (trials 3-5), novel object-related activity significantly increased and matched the activity for familiar object trials.
Read also: Learning Civil Procedure
tags: #doe #learning #nucleus #definition
