書字動作の習熟予測

Prediction of learning in Writing Performance

要旨：作業療法の治療手段として，問題のある実際機能そのものを繰り返し練習させ，技能習熟を図ることは少なくない．作業療法の特長である目的活動の使用にも叶っている．脳卒中患者に目的活動として書字動作訓練がしばしば用いられる．しかし，練習量から学習を予測する研究は余りなされていなかった．今回，延べ8名のCVA患者の100〜500日分の書字訓練の記録を分析して，練習回数と書字速度の改善(書字動作習熟)の関係をしらべ，簡単な数式(習熟モデル)で表現できるか，このモデルによる習熟予測の可能性を調べた．その結果，練習回数と書字動作習熟は両対数グラフ上で直線的に表現でき，習熟現象を簡単な指数関数T_N＝A/N^Bで表せた．この回帰式を用いて3カ月以上先の書字動作を1分内の誤差で予測できる可能性を示した．更にこの学習モデルの臨床応用についても考察した．

　As a treatment modality, occupational therapists utilize activities that patients actually have difficulty in performing in their daily lives. The modality fits the philosophy of using purposeful activities. Writing is one of these purposeful activities for CVA patients with writing difficulties.

　Because there has been little research done on the amount of practice and its effects, the author studies the following questions about learning performance. What kinds of relationship exists between the amount of practice and writing speed? How can the relationships be expressed in a simple mathematical formula? Can this learning model be applied generally in predicting writing performance?

　This author examined 8 CVA subjects for 100 to 500 days of writing practice. The author found that the relationship between the amount of practice and writing speed can be expressed in a linear relationship on the log-log paper and in a simple power formula (T_N=A/N^B). A prediction of 3 months or more can be made for future performance. The author also discussed the clinical application of this learning model.