Selective influence of working memory load on exceptionally slow reaction times

Area of Science:

• Cognitive Psychology
• Neuroscience
• Human Factors

Background:

• Slow reaction times, particularly the long tail of reaction time distributions, are prevalent in populations with cognitive deficits.
• Previous research indicates a strong correlation between working memory (WM) capacity and the magnitude of this long-tail effect.
• The causal relationship between working memory and slow reaction times has remained unclear.

Purpose of the Study:

• To investigate the causal role of working memory (WM) availability in the generation of exceptionally slow reaction times.
• To determine whether WM demands directly influence the parameters of the reaction time distribution, specifically the long tail.
• To differentiate the effects of WM load on decision-making versus non-decision processes like rule retrieval.

Main Methods:

• Conducted three choice-reaction task experiments manipulating working memory (WM) demands.
• Varied stimulus-response rules (arbitrary vs. nonarbitrary, novel vs. practiced, declarative vs. procedural WM load).
• Utilized ex-Gaussian model fitting and evidence accumulation modeling to analyze reaction time distributions and decision processes.

Main Results:

• Working memory (WM) demands significantly influenced the τ parameter of the ex-Gaussian model, which characterizes the long tail of the reaction time distribution.
• WM load primarily affected the non-decision component of reaction time, presumed to reflect rule retrieval, rather than the decision-making process itself.
• The impact of WM demands on slow reaction times was consistent across different experimental manipulations.

Conclusions:

• Working memory (WM) availability plays a crucial causal role in the generation of exceptionally slow reaction times.
• The effect of WM load on slow reaction times is mediated by its influence on rule retrieval processes, not by altering the speed of decision-making.
• Findings suggest that interventions targeting working memory and rule retrieval may help mitigate slow reaction times in various clinical and aging populations.