Refine
Document Type
- Article (3)
- Doctoral Thesis (1)
Language
- English (4)
Has Fulltext
- yes (4)
Is part of the Bibliography
- no (4)
Keywords
- reward (4) (remove)
Institute
- Extern (1)
- Frankfurt Institute for Advanced Studies (FIAS) (1)
- Medizin (1)
- Psychologie (1)
In human neuroscientific research, there has been an increasing interest in how the brain computes the value of an anticipated outcome. However, evidence is still missing about which valuation related brain regions are modulated by the proximity to an expected goal and the previously invested effort to reach a goal. The aim of this dissertation is to investigate the effects of goal proximity and invested effort on valuation related regions in the human brain. We addressed this question in two fMRI studies by integrating a commonly used reward anticipation task in differential versions of a Multitrial Reward Schedule Paradigm. In both experiments, subjects had to perform consecutive reward anticipation tasks under two different reward contingencies: in the delayed condition, participants received a monetary reward only after successful completion of multiple consecutive trials. In the immediate condition, money was earned after every successful trial. In the first study, we could demonstrate that the rostral cingulate zone of the posterior medial frontal cortex signals action value contingent to goal proximity, thereby replicating neurophysiological findings about goal proximity signals in a homologous region in non-human primates. The findings of the second study imply that brain regions associated with general cognitive control processes are modulated by previous effort investment. Furthermore, we found the posterior lateral prefrontal cortex and the orbitofrontal cortex to be involved in coding for the effort-based context of a situation. In sum, these results extend the role of the human rostral cingulate zone in outcome evaluation to the continuous updating of action values over a course of action steps based on the proximity to the expected reward. Furthermore, we tentatively suggest that previous effort investment invokes processes under the control of the executive system, and that posterior lateral prefrontal cortex and the orbitofrontal cortex are involved in an effort-based context representation that can be used for outcome evaluation that is dependent on the characteristics of the current situation.
The intrinsic complexity of the brain can lead one to set aside issues related to its relationships with the body, but the field of embodied cognition emphasizes that understanding brain function at the system level requires one to address the role of the brain-body interface. It has only recently been appreciated that this interface performs huge amounts of computation that does not have to be repeated by the brain, and thus affords the brain great simplifications in its representations. In effect the brain’s abstract states can refer to coded representations of the world created by the body. But even if the brain can communicate with the world through abstractions, the severe speed limitations in its neural circuitry mean that vast amounts of indexing must be performed during development so that appropriate behavioral responses can be rapidly accessed. One way this could happen would be if the brain used a decomposition whereby behavioral primitives could be quickly accessed and combined. This realization motivates our study of independent sensorimotor task solvers, which we call modules, in directing behavior. The issue we focus on herein is how an embodied agent can learn to calibrate such individual visuomotor modules while pursuing multiple goals. The biologically plausible standard for module programming is that of reinforcement given during exploration of the environment. However this formulation contains a substantial issue when sensorimotor modules are used in combination: The credit for their overall performance must be divided amongst them. We show that this problem can be solved and that diverse task combinations are beneficial in learning and not a complication, as usually assumed. Our simulations show that fast algorithms are available that allot credit correctly and are insensitive to measurement noise.
Intrinsic motivation, the causal mechanism for spontaneous exploration and curiosity, is a central concept in developmental psychology. It has been argued to be a crucial mechanism for open-ended cognitive development in humans, and as such has gathered a growing interest from developmental roboticists in the recent years. The goal of this paper is threefold. First, it provides a synthesis of the different approaches of intrinsic motivation in psychology. Second, by interpreting these approaches in a computational reinforcement learning framework, we argue that they are not operational and even sometimes inconsistent. Third, we set the ground for a systematic operational study of intrinsic motivation by presenting a formal typology of possible computational approaches. This typology is partly based on existing computational models, but also presents new ways of conceptualizing intrinsic motivation. We argue that this kind of computational typology might be useful for opening new avenues for research both in psychology and developmental robotics.
Adolescence has been linked to an enhanced tolerance of uncertainty and risky behavior and is possibly connected to an increased response toward rewards. However, previous research has produced inconsistent findings. To investigate whether these findings are due to different reward probabilities used in the experimental design, we extended a monetary incentive delay (MID) task by including three different reward probabilities. Using functional magnetic resonance imaging, 25 healthy adolescents and 22 adults were studied during anticipation of rewards in the VS. Differently colored cue stimuli indicated either a monetary or verbal trial and symbolized different reward probabilities, to which the participants were blinded. Results demonstrated faster reaction times for lower reward probabilities (33%) in both age groups. Adolescents were slower through all conditions and had less activation on a neural level. Imaging results showed a three-way interaction between age group x condition x reward probability with differences in percent signal change between adolescents and adults for the high reward probabilities (66%, 88%) while adolescents demonstrated differences for the lowest (33%). Therefore, previous inconsistent findings could be due to different reward probabilities, which makes examining these crucial for a better understanding of adolescent and adult behavior.