Bad habits-good goals? Meta-analysis and translation of the habit construct to alcoholism

Abstract

Excessive alcohol consumption remains a global public health crisis, with millions suffering from alcohol use disorder (AUD, or simply "alcoholism"), leading to significantly reduced life expectancy. This review examines the interplay between habitual and goal-directed behaviors and the associated neurobiological changes induced by chronic alcohol exposure. Contrary to a strict habit-goal dichotomy, our meta-analysis of the published animal experiments combined with a review of human studies reveals a nuanced transition between these behavioral control systems, emphasizing the need for refined terminology to capture the probabilistic nature of decision biases in individuals with a history of chronic alcohol exposure. Furthermore, we distinguish habitual responding from compulsivity, viewing them as separate entities with diverse roles throughout the stages of the addiction cycle. By addressing species-specific differences and translational challenges in habit research, we provide insights to enhance future investigations and inform strategies for combatting AUD.

Fig. 1. Neurocircuitry and experimental paradigms of striatal learning.

A Left: Corticostriatal projections originate from distinct non-overlapping populations of neurons. Cortical neurons were retrogradely labeled by injection of viral tracers (ssAAV-retro/2-hSyn1-mCherry-WPRE-hGHp and ssAAV-retro/2-hSyn1-EGFP-WPRE-hGHp) within the posterior dorsomedial DMS and anterior DLS. Fluorescence-labeled neurons of the mPFC project to DMS (green), while M1 neurons project to DLS (purple). In the OFC distinct populations are found that project either to DMS or DLS. Scale bar: 1 mm. Middle: Simplified representation of the prefrontocortical input to the dorsal striatum. Neurons from M1 and SMC project to the DLS (purple), mPFC neurons to the DMS (green), and OFC neurons project to both regions (purple and green). Striatal dopaminergic input from the VTA and SNc are shown in blue. Right: Coronal sections of the rodent and human brain showing the main striatal regions with black arrows representing ventromedial to dorsolateral information transfer. The dorsal part of the striatum can be subdivided into the DMS (rodents) and caudate nucleus (humans) (green) and the DLS (rodents) and putamen (humans) (purple). Functional aspects of this circuitry are described in Box 2. B Typical instrumental (operant box), spatial navigation (T-maze), and skilled walking (horizontal ladder) tasks that are used to assess biases of goal-directed or automatic response tendencies in rodents. C Human sequential decision-making task (2-step task) to assess model-based and model-free learning (see Box 1). In each trial, participants perform two sequential decisions at two stages in order to obtain probabilistic monetary rewards. In this version [118], participants start from planet Earth (1st stage) and choose between one out of two rockets, in order to land on one out of two planets (2nd stage), each inhabited by two different aliens. Importantly, the transition from 1st stage choice to the 2nd stage underlies a probabilistic structure: while one rocket flies commonly (70% probability) to the yellow planet and only rarely to the red one (30% probability), the inverse structure is true for the other rocket. In the 2nd stage, participants chose between one out of two aliens in order to obtain a reward. The reward probabilities associated with each 2nd stage alien vary slowly across trials according to Gaussian random walks in order to foster continuous learning across the task. That way the 2-step task allows to dissociate model-based from model-free behavior: While pure model-free control simply increases the choice probability of actions rewarded in previous trials, model-based control additionally considers if rewards followed a common or a rare 2^nd stage transition, i.e., takes into account the underlying task structure. D Schematic learning curve of a behavior. Early into training the behavior may be less accurate or efficient (purple) but will gradually improve. At later time points, performance has stabilized but is less flexible (green) and resistant to interference. Behavioral automaticity such as degree of goal-directed responding can be assessed in early or late phase. Neuroanatomical abbreviations were used according to the rat brain atlas [119]: DLS dorsolateral striatum, DMS dorsomedial striatum, mPFC medial prefrontal cortex, Acb nucleus accumbens, OFC orbitofrontal cortex, M1 primary motor cortex, SMC sensorimotor cortex, SNc substantia nigra pars compacta, VTA ventral tegmental area.

Fig. 2. Meta-analysis on the effect of chronic alcohol pre-exposure on responding in reward devaluation tests in rodents.

The forest plot shows the standardized effect size (Hedges’ g_av) representing the difference in responding between reward-devalued and non-devalued conditions of the alcohol and control groups (17 comparisons) from 10 published studies. The experiments included testing different operant schedules (experiments Nr. 3–7), rewards (exp. Nr. 10), reward-lever contingencies (exp. Nr. 16), or time points (exp. Nr. 17). Blue and red squares represent the control and alcohol conditions, respectively, with their position relative to the x-axis indicating the effect size and their area representing their percent weight within the meta-analysis, based on the variance of the effect size. Horizontal lines indicate the confidence intervals (CI), with the values given in the adjacent table. Colored vertical bars represent the CI of the overall effect of control (blue) and alcohol (red) conditions, respectively, also displayed at the base of the plot. The vertical dashed line represents the zero effect, i.e., no devaluation or full habitual behavior. Study variables are shown to the right and include sex and age (early adolescent—EA, late adolescent—LA, adult—AD), route of alcohol administration (oral intake, intra-peritoneal injection, CIE vapor, shown by symbols), and type of reward (white pellets for sucrose pellets, brown pellets for food pellets, blue drops for sweet solution, yellow drops for alcohol solution, shown in symbols), and test condition (satiety devaluation—SD, contingency degradation—CD). Reward symbols beneath the left lever indicate that a single reward was tested (no choice), while those beneath both levers signify that two rewards were tested simultaneously (choice). In exps. Nr. 7 and 10, two rewards were tested separately (no choice).