1 / 25

범주학습의 feedback 유형과 두뇌 활성화의 차이

범주학습의 feedback 유형과 두뇌 활성화의 차이. 강은주 1 , 이지현 1 , 안신혜 1 , 김성일 2 1 강원대학교 심리학과 , 2 고려대학교 교육학과. Instrumental Conditioning. Learning contingences between its own responses and rewarding or punishing outcome  reinforcement learning (Stutton and Barto, 1998)

savannah-walker
Download Presentation

범주학습의 feedback 유형과 두뇌 활성화의 차이

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. 범주학습의 feedback유형과 두뇌 활성화의 차이 강은주1, 이지현1, 안신혜 1, 김성일2 1강원대학교 심리학과, 2고려대학교 교육학과

  2. Instrumental Conditioning • Learning contingences between its own responses and rewarding or punishing outcome  reinforcement learning (Stutton and Barto, 1998) • Two processes (.. O’Doherty et al., 2004 Science) • Critics: use a temporal difference prediction error signal to update successive predictions of future reward associated with being at a state of the external and internal environment (determined by the arrangement of stimuli) • Actors: use a similar signal to modify stimulus-response or stimulus-response-reward associations in the from of policy, so that actions associated with greater long-term reward are chosen more frequently on subsequent trials Functional NeuroImaging Lab

  3. Neural correlate of these temporal difference prediction error signal • the phasic activity of dopamine neurons  projections to the ventral and dorsal striatum • Ventral striatum : reward and motivation • Dorsal striatum: motor and cognitive control Functional NeuroImaging Lab

  4. Methods • Volunteers • 22 normal volunteers (mean 22.7 yr, 16F/6M) • fMRI data acquisition • 1.5T MRI, Phillips, SENSE-EPI, TR = 3s, TE = 50ms, flip angle = 90˚, FOV = 22cm, 22 5mm slices, 180 scans • Paradigm • 4 repetitions of 5 alternating conditions/ run x 3 • Fixation, 1 pseudo conditioning (Dummy), 3 learning conditions depending on feedback type • a categorical visuo-motor learning task where either left or right hand button-pressing response for a visually presented English consonant letter (cue, 2.5s duration) was followed by a feedback. • Analysis • SPM5 mixed design, event types, block design Functional NeuroImaging Lab

  5. Feedback types Functional NeuroImaging Lab

  6. Feedback Type and Correct Response Condition x Session P<.001 Functional NeuroImaging Lab

  7. Feedback Type and Response Time Functional NeuroImaging Lab

  8. Dummy Neutral Negative Positive

  9. Learning. = dummy > Fixation Medial Superior PFC region MNI = -6, 28, 46 (BA8) uP<.0001, Caudate Body MNI = -4,-16,24 uP<.0001, Parietal region MNI = 32, -62, 42, DLPFC MNI = 30, 6, 62 uP<.0001, Functional NeuroImaging Lab

  10. Learning > Fixation Positive Neutral Negative Midbrain region MNI = -12, -12, -14, uP<.0001, Functional NeuroImaging Lab Local maxima cP<.05

  11. Learning > Dummy L Left Posterior Hippocampus MNI = -34, -44, -2 uP<.0001 Dorsal Hypothalamus MNI = -2, 0, -8, uP<.0001

  12. Dummy > Learning L L DLPFC MNI = -40, 12, 52 uP<.0001 middle temporal gyrus MNI = -34, -44, -2 uP<.0001 Cerebellum MNI = 10, -80, -32 uP<.0001 Insula MNI = -28, 26, -4 uP<.0001 Local maxima cP<.05

  13. Hit > Miss R Putamen MNI = -28, 4, 4 N.Acc MNI = -16, 12, -12 uP<.001 Functional NeuroImaging Lab

  14. Hit Miss R uP<.0001 Functional NeuroImaging Lab

  15. 6 feedback events Functional NeuroImaging Lab

  16. Reward > Punishment Ventral Striatum/N.Acc MNI = 10, 16, -6 cP<.05 Functional NeuroImaging Lab

  17. NAC: activation or deactivation? R Activations to Punishment Activations to Reward Deactivations To Punishment Deactivations to Reward Ventral Striatum/N.Acc MNI = 10, 16, -6 uP<.0001 Functional NeuroImaging Lab

  18. Correct & Incorrect R Activationsto Correct Activations to Incorrect Functional NeuroImaging Lab

  19. Reward & Dummy_Reward R Reward both Dummy Reward y=-12 X=4 X=0 uP<.0001 (T=4.49) Mammillothalamic Tract Activations Functional NeuroImaging Lab

  20. Functional NeuroImaging Lab

  21. Correct Feedback> Incorrect Feedback R Ventral Striatum/N.Acc MNI = -18, 2, -12, T=5.85 uP<.0001 Functional NeuroImaging Lab

  22. Positive Feedback> Negative Feedback R Ventral Striatum MNI = 10, 14, -6, T=8.08 Posterior Cingulate MNI = 2,-44,32, T=7.67 N.Acc MNI = -22, 4, -12, T=12.17 Positive Feedback Negative Feedback Positive > Negative cP<.05 Functional NeuroImaging Lab

  23. Limbic corticostrial loop and relationship with other cortex Functional NeuroImaging Lab Cardinal et al., 2002

  24. Conclusions • Circuits for instrumental learning • Conditioning vs. Pseudo-Learning • P. Hippocampus, dorsal hypothalamus • Dorsal striatum vs. ventral striatum • Feedback events  N.Acc/Ventral striatum, Hypothalamus, Parietal Cortex, DLPFC, Posterior Hippocampus • Individual Difference • VTA only for reward learning Functional NeuroImaging Lab

  25. Individual Difference during Positive Condition VTA MNI = 0 -16 -22 uP<.01 Functional NeuroImaging Lab

More Related