RÉGUIGNE-KHAMASSI Mehdi
Supervision : Agnès GUILLOT
Co-supervision : WIENER Sidney
Complementary roles of the rat prefrontal cortex and striatum in reward-based learning and shifting navigation strategies
In order to efficiently reach resources, mammals are able to behave according to different navigation "strategies", and can shift from one to the other when environmental contingencies change. Stimulus-Response (S-R) strategies are usually distinguished from more complex strategies requiring the elaboration of a world model. This thesis employs a multidisciplinary approach – neurophysiology, computational modeling, and simulated robotics – to study the roles of the rat medial prefrontal cortex (mPFC) and striatum in learning and shifting strategies. The present work suggests that the rat ventral striatum shows a reward anticipation activity coherent with the « Actor-Critic » theory proposed in dynamic systems studies, and which can be combined with self-organizing maps to enable a simulated robot to solve a S-R strategy. The mPFC seems to play a role in the detection of environment changes, in the detection of drops in the animal's performance, and in the selection of a more appropriate strategy.
Defence : 09/26/2007
Jury members :
Pr. Alain Berthoz, Examinateur, LPPA, Collège de France
Pr. Philippe Bidaud, Examinateur, ISIR, Université Paris 6
Pr. Kenji Doya, Rapporteur, IRP, Okinawa Institute of Sci. and Tech.
Dr. Agnès Guillot, Directrice de thèse, LIP6/ISIR, Université Paris X
Pr. Cyriel Pennartz, Examinateur, SILS, Universiteit van Amsterdam
Dr. Bruno Poucet, Rapporteur, LNC, CNRS-Université de Provence
Dr. Sidney Wiener, Directeur de thèse, LPPA, CNRS-Collège de France
2004-2018 Publications
-
2018
- N. Aklil, B. Girard, L. Denoyer, M. Khamassi : “Sequential Action Selection and Active Sensing for Budgeted Localization in Robot Navigation”, International Journal of Semantic Computing, vol. 12 (01), pp. 109-127, (World Scientific) (2018)
-
2017
- N. Aklil, B. Girard, M. Khamassi, L. Denoyer : “Sequential Action Selection for Budgeted Localization in Robots”, IEEE Robotic Computing 2017, Taichung, Taiwan, Province of China, pp. 97-100 (2017)
-
2006
- S. Wiener, M. Khamassi, A. Peyrache, V. Douchamps, P. Tierney, F. Battaglia : “Transitions in behaviorally correlated activity in medial prefrontal neurons of rats acquiring and switching strategies in a Y-maze”, Society for Neuroscience Abstracts, Atlanta, United States (2006)
- M. Khamassi, L.‑E. Martinet, A. Guillot : “Combining self-organizing maps with mixtures of experts: Application to an Actor-critic model of reinforcement learning in the Basal Ganglia”, From Animals to Animats 9, vol. 4095, Lecture Notes in Computer Science, Rome, Italy, pp. 394-405, (Springer Berlin Heidelberg) (2006)
-
2005
- M. Khamassi, L. Lachèze, B. Girard, A. Berthoz, A. Guillot : “Actor-critic models of reinforcement learning in the basal ganglia: From natural to artificial rats”, Adaptive Behavior, vol. 13 (2), pp. 131-148, (SAGE Publications) (2005)
- J.‑A. Meyer, A. Guillot, B. Girard, M. Khamassi, P. Pirim, A. Berthoz : “The Psikharpax project: Towards building an artificial rat”, Robotics and Autonomous Systems, vol. 50 (4), pp. 211-223, (Elsevier) (2005)
-
2004
- D. Filliat, B. Girard, A. Guillot, M. Khamassi, L. Lachèze, J.‑A. Meyer : “State of the artificial rat Psikharpax”, From Animals to Animats 8, Los Angeles, CA, United States, pp. 3-12, (The MIT Press) (2004)
- M. Khamassi, B. Girard, A. Berthoz, A. Guillot : “Comparing three critic models of reinforcement learning in the basal ganglia connected to a detailed actor in a S-R task”, Intelligent Autonomous Systems 8, Amsterdam, Netherlands, pp. 430-437, (IOS Press) (2004)