Skip to main content
European Commission logo print header

Thinking Ahead: human planning from a predictive processing perspective

Periodic Reporting for period 2 - ThinkAhead (Thinking Ahead: human planning from a predictive processing perspective)

Reporting period: 2021-07-01 to 2022-12-31

Within the ThinkAhead project, we study how people solve planning problems, either individually or collaboratively. To this aim, we design experiments that require adult human participants to plan ahead and we realize computational models of planning tasks that link behavioral evidence to formal principles.
Our studies help shed light on one of the more advanced cognitive abilities of our predictive brains and to address some fundamental questions, such as how humans address the uncertainties related to task representation, what plan to pursue, and how to collaborate with others to solve cooperative planning tasks.
During the project, we perform various human planning experiments that ask how people form internal representations of planning tasks and what are the mechanisms that they use to successfully solve challenging problems (either alone or together with others). We are releasing our experiments as game-like apps that can be freely downloaded and played on mobile phones and other devices, in order to collect behavioral and kinematic data from a diverse set of participants and to face the current problems of conducting in-person experiments.

In parallel, we develop computational models of hierarchical inference and planning that we use to better understand the computational principles that guide human planning and to conduct model-based analyses of participant’s behavior during the experiments.

Finally, we integrate our novel findings on planning within a larger, inferential perspective of how the brain learns internal models of the world and of the body within it - and how it uses these models to address novel and unforeseen real world problems. For this, we are investigating the possible neuronal underpinnings of inferential dynamics, both when the brain is engaged in a particular planning or problem solving task, and when it is at rest – as evident for example in hippocampal “replays” and brain resting state dynamics.
Each of our experiments and models is providing novel insights about the cognitive mechanisms that allow us to solve challenging planning problems that cannot be solved by (mentally) searching exhaustively across all the possible solutions. For example, we found evidence suggesting that people consider the availability of future options and not just future rewards during planning; that they account for cognitive costs when finding shortcuts during navigation; and that during cooperative planning they develop strategies that make them more predictable and easier to monitor. By the end of the project, we aim to develop and validate empirically a comprehensive theory of human planning and link it to comprehensive views of the brain as a prediction machine.