➡️P3.I.45: Validating predictions of a flexible decision-making model for varying decision goals and choice set properties by Ana Hernandez at 11:15 on 10/5.

(2/2)

Our findings demonstrate the critical role that cognitive dynamics play in explaining the mechanisms through which cognitive inflexibility arises in older adulthood.

With increasing age, people move slower through the space of control configurations that determine performance. We also show that the ability to adjust control configurations and the ability to maintain performance despite goal switches is maintained across the lifespan.

Using computational modeling and building on our previous work (psycnet.apa.org/record/2026-...), we measured changes in two control signals (attentional focus and response caution) as people of different ages switched between goals that induced distinct control configurations.

We propose that the speed of movement between control limits cognitive flexibility in older adults. To test this, we had people across the lifespan perform a cognitively demanding task with changing performance goals (perform the task quickly vs. accurately).

Changing goals require adjustments of cognitive control configurations (e.g., level of attentional focus), even within similar tasks (e.g., emailing a friend vs. your boss). We formalize such adjustments as a dynamical system moving from its current state to the new target state.

P.S. Free version of the article is here: www.biorxiv.org/content/10.1...
Check Berkeley’s recent coverage of the work here: ls.berkeley.edu/news/psychol...

Overall, we show that changing control states (attention and caution) to meet a new goal induces control adjustment costs, and that these costs arise from cognitive control dynamics. Good luck with your post-Twitter-scroll goals!

We also show (Study 4) that the frequency of performance goal changes parametrically increases the costs, and that the expectation about change frequency determines the cost.

We also confirmed 2 other predictions our model makes: we show that people exhibit larger costs when target control states are more distant (Study 2) and when they have less time to adjust control (Study 3).

Confirming the prediction of the model, in Study 1 we found that control states (defined by levels of threshold and drift rate) are pulled closer together in blocks which demand control adjustments that produce costs.

Due to the time it takes to adjust control states, the model predicts the existence of a control adjustment cost. When frequently moving between different goals (Varying blocks) people will undershoot their target control state.

We develop a dynamical systems model to describe such adjustments in continuous control signals. Our model proposes that control states are adjusted gradually from their current state toward the target state specified by the new performance goal.

Different performance goals require different cognitive control states. Performing a task quickly can be done with low levels of caution (Threshold) and attention (Drift rate), but being accurate (Accuracy goal) requires an increase in caution and attention.