Different physical and mental conditions cause the human heart rate to fluctuate. However, it is difficult to record physiological data in real sports situations without impeding an athlete's movement, and no effective method has been established for quantifying changes in an athlete's heart rate that are unrelated to exercise intensity. Using a recently developed wearable bioelectrode ("hitoe"), we performed electrocardiographic and bodily acceleration measurements on two baseball players during actual practice and game situations. Analyzing the practice data, we constructed a linear regression model that accurately predicts each player's current heart rate from the previous 15-min acceleration data (R^2=0.86). Using this model and recorded acceleration data, we estimated the heart rates of the baseball players during a game. The results showed that the recorded heart rate deviated rapidly from the estimated heart rate relating to some game situations. Therefore, our proposed method effectively detected increases in the players' heart rates that were related to the game situations rather than to the exercise intensity, which could be interpreted as heart rate changes induced by the players' mental state.
