- Email: [email protected]
Two-stage learning in eyeblink conditioning in mice
S22
Abstracts / Neuroscience Research 58S (2007) S1–S244
S3A-E4 Two-stage learning in eyeblink conditioning in mice
S3A-F3 The anterior cingulate cortex in learning and reward-
Shigeyoshi Itohara, Hye-Soo Kim, Jin-Sung Park Behavioral Genetics, RIKEN BSI, Wako, Japan
guided decision making
Delay eyeblink conditioning is one of the most extensively characterized paradigms for motor learning and memory. Cumulative evidence indicates an essential role for the cerebellum in this paradigm. The site and mechanisms underlying memory formation, however, are under debate. The anterior interpositus nucleus (AIN) is a proposed site for the association memory. To unravel the mechanisms underlying memory formation, we systematically examined transcripts that change during conditioning in mice. There were two groups of genes with distinct spatial and temporal characteristics. Representative EARLY gene expression peaked on 1 d of training in broad cerebellar areas in both the paired and unpaired conditioning groups. Representative LATE gene expression was selectively increased in the AIN of the 7-d paired group, but not the unpaired group. These data fit well with the two-stage learning theory, which proposes emotional and motor learning phases, and support the role for the AIN as a memory site. Using EARLY genes as molecular probes, we are now investigating the role of emotional learning in eyeblink conditioning. Research fund: KAKENHI18022049
S3A-E5 Cerebellar plasticity at ultrastructural level Ryuichi Shigemoto Division of Cerebral Structure, National Institute for Physiological Sciences, Okazaki, Japan Morphological substrates for the memory formation in vivo have not been well demonstrated. We used short- and long-term adaptation of horizontal optokinetic response (HOKR) to study the link between learning and underlying ultrastructual changes in synapses and synaptic AMPAtype glutamate receptors. We used SDS-digested freeze-fracture replica labeling, which can detect AMPA receptor channels with almost one-toone detection sensitivity. Almost all PF-PC synapses were labeled for AMPA receptors with a large variability in number and density of particles. In short-term adaptation of HOKR induced with one hour training, we detected a significant reduction of AMPA receptors by 25% at PFPC synapses in the flocculus. On the contrary, in long-term adaptation induced with the one hour training for 5 consecutive days, we detected no significant changes in AMPA receptor density but found a significant reduction of synapse density by 33% in the flocculus. These results suggest that memory traces for short- and long-term memory are differently formed in the cerebellar motor learning.
Matthew F.S. Rushworth 1,2 , Timothy E.J. Behrens 1,2 , Mark E. Walton 1 1 Department of Experimental Psychology, University of Oxford, Oxford, England, UK; 2 FMRIB, University of Oxford, Oxford, England, UK Signal changes in the human anterior cingulate cortex (ACC) can be recorded when people make choices and when the outcomes of the choices will be used to guide subsequent behaviour. The size of the ACC signal varies with the degree to which the choice is free and unconstrained and with the importance of the outcome as a determinant for future behaviour regardless of whether or not the outcome is positive or negative. A recent fMRI study suggests that ACC activity changes as the length of the reward history that is needed to determine the next choice changes. Such information is critical for setting the rate at which learning occurs. When lesions are made in the ACC, macaques are as sensitive as controls to a decrement in reinforcement but the influence of the extended reward history on choice is diminished. Research funds: MRC, Royal Society
S3A-F4 Top-down attention to prediction errors of action values in the prefrontal cortex Kenji Matsumoto, Madoka Matsumoto, Hiroshi Abe, Keiji Tanaka RIKEN Brain Science Institute, Saitama, Japan When we are not sure of appropriate action in a given situation, we attend to the outcome of tried action to know its appropriateness for the next time. To uncover neural processes underlying such top-down attention in action selection learning, we recorded the neuronal activity in the medial and lateral prefrontal cortex (PFC) in the monkeys performing a task in which learning of action selection was repeated with visual feedback signals. About 15% of neurons in both the medial and lateral PFC gradually increased the firing toward the visual feedback onset in early phases of learning, and the activity disappeared after the appropriate action became certain. More exactly, the probability-weighted average of absolute discrepancy between the predicted value and actual outcome of potential actions gave the best fit of the changes in magnitude of the neuronal activity along learning. In each trial, the activity appeared earlier in the medial PFC than the lateral PFC. These results suggest that the medial PFC is essential in the top-down attention in action selection learning.
Research funds: SORST, Japan Science and Technology Corporation
S3A-F1 Action value in the striatum and reinforcementlearning model of cortico-basal ganglia network Kazuyuki Samejima 1 , Yasumasa Ueda 2 , Kenji Doya 3 , Minoru Kimura 2 1 Tamagawa University Brain Science Institute, Tokyo, Japan; 2 Department of Physiology, Kyoto Prefectural University of Medicine, Kyoto, Japan; 3 IRP, OIST, Okinawa, Japan When we face a choice, evaluating the consequences of the choice is important. It has been suggested that the basal ganglia play a crucial role in reward expectation and voluntary movement. But exactly how the basal ganglia contribute to the processing between reward expectation and decision-making remains unclear. To address this issue, we recorded monkey striatal projection neurons during a free-choice task in which two monkeys turned a handle to either left or right direction. In the delay period before the choices, magnitude of discharge rate in majority of recorded neurons (53/142) was correlated with the values of one of the two actions, which were estimated by a reinforcement learning (RL) model to fit history of monkey’s choices and outcomes. These results supported a RL model of basal ganglia, in which action values are represented in the striatum, compared between possible actions to select one, and updated by expectation errors between the selected action value and obtained reward. Research funds: KAKENHI (17022032, 1800383, 18019033)
S3A-G2 Microglial activation by TLR ligands: Usual and unusual suspects Uwe-Karsten Hanisch Institute of Neuropathology, University of Goettingen, Goettingen, Germany To sense environmental signals, microglial cells are equipped with an array of receptors. They also express several Toll-like receptors (TLR) and mount diverse responses to their agonists. We focussed on a protein tyrosine kinase activity which is involved in the signalling of various TLR. The PTK assists the MyD88 route and controls the intracellular calcium, thereby affecting activation-associated functions. Moreover, TLR may serve the recognition of a broader variety of signals indicating homeostatic disturbance. We identified endogenous proteins which apparently activate microglia through a TLR4/MyD88-dependent mechanism. LPS, the prototypic microbial ligand of TLR4, can be ruled as a confounding factor. The ability of the ligands to trigger TLR4 signalling seems to depend on critical structural requirements. Interestingly, although the ligands use TLR4 for mandatory signalling, additional (receptor) components appear to fine-tune respective consequences. TLR4 may thus serve responses to different microglia-challenging stimuli but the induced responses distinguish between individual agonists. Research fund: Hertie Foundation
Abstracts / Neuroscience Research 58S (2007) S1–S244
S3A-E4 Two-stage learning in eyeblink conditioning in mice
S3A-F3 The anterior cingulate cortex in learning and reward-
Shigeyoshi Itohara, Hye-Soo Kim, Jin-Sung Park Behavioral Genetics, RIKEN BSI, Wako, Japan
guided decision making
Delay eyeblink conditioning is one of the most extensively characterized paradigms for motor learning and memory. Cumulative evidence indicates an essential role for the cerebellum in this paradigm. The site and mechanisms underlying memory formation, however, are under debate. The anterior interpositus nucleus (AIN) is a proposed site for the association memory. To unravel the mechanisms underlying memory formation, we systematically examined transcripts that change during conditioning in mice. There were two groups of genes with distinct spatial and temporal characteristics. Representative EARLY gene expression peaked on 1 d of training in broad cerebellar areas in both the paired and unpaired conditioning groups. Representative LATE gene expression was selectively increased in the AIN of the 7-d paired group, but not the unpaired group. These data fit well with the two-stage learning theory, which proposes emotional and motor learning phases, and support the role for the AIN as a memory site. Using EARLY genes as molecular probes, we are now investigating the role of emotional learning in eyeblink conditioning. Research fund: KAKENHI18022049
S3A-E5 Cerebellar plasticity at ultrastructural level Ryuichi Shigemoto Division of Cerebral Structure, National Institute for Physiological Sciences, Okazaki, Japan Morphological substrates for the memory formation in vivo have not been well demonstrated. We used short- and long-term adaptation of horizontal optokinetic response (HOKR) to study the link between learning and underlying ultrastructual changes in synapses and synaptic AMPAtype glutamate receptors. We used SDS-digested freeze-fracture replica labeling, which can detect AMPA receptor channels with almost one-toone detection sensitivity. Almost all PF-PC synapses were labeled for AMPA receptors with a large variability in number and density of particles. In short-term adaptation of HOKR induced with one hour training, we detected a significant reduction of AMPA receptors by 25% at PFPC synapses in the flocculus. On the contrary, in long-term adaptation induced with the one hour training for 5 consecutive days, we detected no significant changes in AMPA receptor density but found a significant reduction of synapse density by 33% in the flocculus. These results suggest that memory traces for short- and long-term memory are differently formed in the cerebellar motor learning.
Matthew F.S. Rushworth 1,2 , Timothy E.J. Behrens 1,2 , Mark E. Walton 1 1 Department of Experimental Psychology, University of Oxford, Oxford, England, UK; 2 FMRIB, University of Oxford, Oxford, England, UK Signal changes in the human anterior cingulate cortex (ACC) can be recorded when people make choices and when the outcomes of the choices will be used to guide subsequent behaviour. The size of the ACC signal varies with the degree to which the choice is free and unconstrained and with the importance of the outcome as a determinant for future behaviour regardless of whether or not the outcome is positive or negative. A recent fMRI study suggests that ACC activity changes as the length of the reward history that is needed to determine the next choice changes. Such information is critical for setting the rate at which learning occurs. When lesions are made in the ACC, macaques are as sensitive as controls to a decrement in reinforcement but the influence of the extended reward history on choice is diminished. Research funds: MRC, Royal Society
S3A-F4 Top-down attention to prediction errors of action values in the prefrontal cortex Kenji Matsumoto, Madoka Matsumoto, Hiroshi Abe, Keiji Tanaka RIKEN Brain Science Institute, Saitama, Japan When we are not sure of appropriate action in a given situation, we attend to the outcome of tried action to know its appropriateness for the next time. To uncover neural processes underlying such top-down attention in action selection learning, we recorded the neuronal activity in the medial and lateral prefrontal cortex (PFC) in the monkeys performing a task in which learning of action selection was repeated with visual feedback signals. About 15% of neurons in both the medial and lateral PFC gradually increased the firing toward the visual feedback onset in early phases of learning, and the activity disappeared after the appropriate action became certain. More exactly, the probability-weighted average of absolute discrepancy between the predicted value and actual outcome of potential actions gave the best fit of the changes in magnitude of the neuronal activity along learning. In each trial, the activity appeared earlier in the medial PFC than the lateral PFC. These results suggest that the medial PFC is essential in the top-down attention in action selection learning.
Research funds: SORST, Japan Science and Technology Corporation
S3A-F1 Action value in the striatum and reinforcementlearning model of cortico-basal ganglia network Kazuyuki Samejima 1 , Yasumasa Ueda 2 , Kenji Doya 3 , Minoru Kimura 2 1 Tamagawa University Brain Science Institute, Tokyo, Japan; 2 Department of Physiology, Kyoto Prefectural University of Medicine, Kyoto, Japan; 3 IRP, OIST, Okinawa, Japan When we face a choice, evaluating the consequences of the choice is important. It has been suggested that the basal ganglia play a crucial role in reward expectation and voluntary movement. But exactly how the basal ganglia contribute to the processing between reward expectation and decision-making remains unclear. To address this issue, we recorded monkey striatal projection neurons during a free-choice task in which two monkeys turned a handle to either left or right direction. In the delay period before the choices, magnitude of discharge rate in majority of recorded neurons (53/142) was correlated with the values of one of the two actions, which were estimated by a reinforcement learning (RL) model to fit history of monkey’s choices and outcomes. These results supported a RL model of basal ganglia, in which action values are represented in the striatum, compared between possible actions to select one, and updated by expectation errors between the selected action value and obtained reward. Research funds: KAKENHI (17022032, 1800383, 18019033)
S3A-G2 Microglial activation by TLR ligands: Usual and unusual suspects Uwe-Karsten Hanisch Institute of Neuropathology, University of Goettingen, Goettingen, Germany To sense environmental signals, microglial cells are equipped with an array of receptors. They also express several Toll-like receptors (TLR) and mount diverse responses to their agonists. We focussed on a protein tyrosine kinase activity which is involved in the signalling of various TLR. The PTK assists the MyD88 route and controls the intracellular calcium, thereby affecting activation-associated functions. Moreover, TLR may serve the recognition of a broader variety of signals indicating homeostatic disturbance. We identified endogenous proteins which apparently activate microglia through a TLR4/MyD88-dependent mechanism. LPS, the prototypic microbial ligand of TLR4, can be ruled as a confounding factor. The ability of the ligands to trigger TLR4 signalling seems to depend on critical structural requirements. Interestingly, although the ligands use TLR4 for mandatory signalling, additional (receptor) components appear to fine-tune respective consequences. TLR4 may thus serve responses to different microglia-challenging stimuli but the induced responses distinguish between individual agonists. Research fund: Hertie Foundation