Building momentum: A computational account of persistence toward long-term goals

Abstract

Additional Information

Author summary: Goals take time to accomplish and require commitment over extended periods of time. However, over time, some goals may become less attainable than others, necessitating switching between goals as contexts and circumstances change. A fundamental question concerns how humans switch goals across extended intervals. One possibility is that humans prospectively commit to specific goals in a way that is maximally sensitive to the likelihood of achieving that goal in the near future. Alternatively, humans might retrospectively persist in goals that they have previously worked toward. Consistent with this latter possibility, we found evidence that humans are retrospectively biased towards goals that they have spent time building progress in, even when it is more optimal to switch. We account for such a preference for accrued progress using a computational model incorporating the concept of momentum borrowed from classical mechanics. We show that momentum steadily integrates reinforcement throughout the goal progress and maintains stable goal commitment toward the goal despite environmental shifts. Although this results in suboptimal performance in our experimental paradigm, we show that momentum computations for goal commitment can have adaptive advantages for goal pursuit in real-world scenarios.

Copyright and License

© 2025 Aenugu, O’Doherty. This is an open access article distributed under the
terms of the Creative Commons Attribution License, which permits unrestricted use,
distribution, and reproduction in any medium, provided the original author and source are credited.

Acknowledgement

We thank Antonio Rangel, Colin Camerer, Caltech SDN graduate program students, and O’Doherty lab members for their valuable feedback and support throughout the development of the manuscript.

Funding

This work was supported by a grant from the ARO MURI project W911NF2110328 to JO’D. SA and JO’D have received salary support from this funding. SA’s salary is also supported by the Tianqiao & Chrissy Chen Center for social and decision neuroscience. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

Data Availability

Data and code used in the study is available at https://github.com/asneha213/goal-switching/tree/main.