A Method of Generating a Dialogue Pattern to Induce Awareness Based on a Reflection Support Agent
Patients with acquired brain injuries often present with adverse symptoms, such as attention, memory, and functional disorders, as well as aphasia, which prevent them from effectively executing activities of daily living. Cognitive rehabilitation can be conducted with these patients to mitigate these issues. A crucial component for the efficacy of this rehabilitation is the patient’s self-awareness. To induce “awareness”, patients are exposed to a process called reflection dialogue, wherein they watch a video of themselves interacting with a cognitive rehabilitation specialist. This process allows the patient to objectively observe themselves. However, this reflection dialogue process requires appropriate specialization according to the symptoms and needs of the patient, which depends on the experience of the cognitive rehabilitation specialist.
The current study aimed to automatize this process using a system that includes a home robot that acts as an agent during the reflection dialogue. This system included three components: (1) Generation of a reflection dialogue; (2) Generation of agentive behavior, based on reinforcement learning of patients’ symptoms (according to patients’ daily logs); and (3) Generation of an appropriate conversation interval (i.e., timing/tempo of the human-robot conversation). This research used conversation methods in a reflection dialogue to induce awareness. Specifically, the current study proposed a method of generating a dialogue pattern through interactions with a reflection support agent, based on a Bayesian network that utilizes communication history. Results show the possibility of using a robot and web-agent to create dialogue. Further, it was found that the “Personal” type of reflection dialogue was significantly better at noticing remarks than the “General” type of reflection dialogue.
KeywordsReflection dialogue Awareness Bayesian network Agent Cognitive rehabilitation
This work was supported by JSPS KAKENHI Grant Number 15K00368. We would like to thank Editage (www.editage.jp) for English language editing.
- 1.Itadume, I.: Disorder Psychology of Acquired Brain Injury. Gakubunsha in Japan (2007)Google Scholar
- 2.Honda, T.: The Rehabilitation of the Patient with Acquired Brain Injury. Igaku-Shoin in Japan (2006)Google Scholar
- 3.Tokyo Metropolitan Government Brain Dysfunction Investigation Committee: Survey Report on the Actual Situation of Higher Brain Dysfunction in Tokyo (2008)Google Scholar
- 7.Ogura, I., et al.: Cooking training for a patient with higher brain dysfunction. J. Cogn. Rehabil. 2007, 40–45 (2007)Google Scholar
- 9.Yamakura, T., Yamazato, M., Inoue, H., Ikejima, C., Asada, T.: Group therapy for individuals with higher brain dysfunction. Cognitive Rehabilitation (2010)Google Scholar
- 10.Tategami, S.: Rusk Institute of Rehabilitation Medicine, Brain Injury Day Treatment Program. Medical School (2010)Google Scholar
- 12.Senzaki, F.: Cognitive rehabilitation for higher brain dysfunction. Cooperative Medical Publisher (2005)Google Scholar
- 13.Nagano, T.: Awareness of higher brain dysfunction. Higher Brain Funct. Res. (2012)Google Scholar
- 14.Meguro, T., Higashinaka, R., Minami, Y.: Dialogue control of listener dialog system using POMDP. The Association for Natural Language Processing (2011)Google Scholar
- 16.Sano, M., Kotani, R., Nakagawa, A., Morimoto, A., Yoshida, Y., Yoshinaga, C.: Dialogue strategy for ADL reflection and cognitive rehabilitation support system using a counseling robot. Hum. Comput. Interact. (2017)Google Scholar
- 17.Adachi, N.: Mind estimation based on communication-robot interaction for cognitive rehabilitation support. Osaka Institute of Technology (2010)Google Scholar
- 18.Murakami, T., Suyama, A., Orihara, R.: Consumer behavior modeling using Bayesian networks (2004)Google Scholar
- 19.Hara, K., Takahashi, K., Ueda, H.: Student behavior modeling and learning using Bayesian networks from questionnaires. Trans. Inf. Process. Jpn. 51(4), 1215–1226 (2010)Google Scholar
- 20.NARA Institute of Science and Technology: CaboCha-NARA Institute of Science and Technology. https://taku910.github.io/cabocha/. Accessed 30 Jan 2019
- 21.Kikuchi, A.: Notes on the researches using KiSS-18. Bull. Fac. Soc. Welfare Iwate Prefectural Univ. 6(2), 41–51 (2004)Google Scholar
- 22.Goldstein, A.P.: Social skill training through structured learningGoogle Scholar
- 23.NTT Data Trusted Global Innovator: BayoLink Bayesian network construction support system. http://www.msi.co.jp/bayolink/. Accessed 30 Jan 2019