Characterizing habit formation in modern slot machine gambling

Wednesday, 23 November, 2022 - 10:50 to 12:20


Habit formation has emerged as a key process in contemporary neuroscience-based models of addictive behaviours, but has received limited attention in the context of gambling and disordered gambling. This study sought to characterize behavioural changes over 3 short sessions of slot machine gambling, with a focus on speed of gambling and flexibility in the betting strategy.

Sixty participants with no prior slot machine experience attended three sessions spaced 6 to 8 days apart, during which they played 200 spins on a realistic simulation of a modern multi-line slot machine (32,112 trials in total). The game was run inside an authentic slot machine chassis with button panel control. Behavioral data were analyzed to characterize habit formation. Fixed-effects regressions assessed predictors of spin initiation latency and bet switches, within and between sessions. Participants gambled faster and became more fixed in their betting strategy as a function of the trial number (latency: B = -0.0024, 95%CI [-0.0025, -0.0023], p < .001; bet switches: OR = 0.995, 95%CI [0.994, 0.995], p < .001) and Session (latency: B = -0.22, 95%CI [-0.23, -0.22], p < .001; bet switches: OR = 0.723, 95%CI [0.698, 0.748], p < .001). In models that distinguished winning from losing trials, we observed behavioural effects of win size and losing streak length. Notably, the ‘post-reinforcement pause’, a slowing following wins compared to losses, increased over sessions (Session x Win Size B = 0.011, 95%CI [0.0026, 0.019], p = .011).

The speed of gambling and rigidity of betting style are highly sensitive to practice both within and across sessions of slot machine gambling, and align well with cognitive criteria for habitual control. The lengthening (i.e. slowing) of the post reinforcement pause with practice indicates that as gambling becomes faster, it also becomes more tightly coupled with contextual cues, in line with conditioning effects.


Presentation files

23 5B 1050 Luke Clark_v1.0.pdf2.03 MBDownload



Part of session