介護食の食感改善のための疑似咀嚼音生成手法に関する基礎検討

摘要

To improve the texture of nursing care foods, we have been developing a pseudo-chewing-sound generator referring to electromyogram (EMG) signals of mastication muscles. Usage of EMG signals enables us to promptly present care recipients a pseudo chewing sound with the intensity corresponding to mastication strength. However, actual chewing sounds vary more than EMG signals in the intensity and duration. In this study, we investigated changes in EMG signals and actual chewing sounds during the mastication of some foods (rice crackers and Japanese pickles) to improve our previously proposed pseudo-chewing-sound generator. As a result, we found that the intensity and duration of the chewing sounds decreased as the number of strokes increased. Furthermore, the intensity and duration of the chewing sounds and the latency between the onset of the EMG signals and that of the chewing sounds were also different between foods. For the EMG signals, only the intensity differed between foods. These effects about the chewing sounds showed statistically large effect sizes according to Cohen's criteria. Based on our findings, we modified the pseudo-chewing-sound generator to enable the intensity and duration of pseudo chewing sounds changeable by the number of mastication strokes and by foods.%我々は介護食の食感を改善することを目的として，咬筋筋電図を変調信号として連続咀嚼音を振幅変調することによる疑似咀嚼音呈示装置を開発してきている.しかし，嚙みしめ状態などでは実際の咀嚼音を模擬できていなかった.本研究では，咀嚼時の咬筋の筋電図と実際の咀嚼音を同時計測し，そのデータに基づいて咬筋筋電図から疑似咀嚼音を生成する方法を見直した.その結果，食物によって振幅に違いは生じるものの咬筋筋電図が咀嚼回数によらずに安定して発生しており，その一方で実際の咀嚼音は食物や咀嚼回数に応じて振幅や持続時間がダイナミックに変化していることが分かった.この知見に基づいて，咬筋筋電図を咀嚼タイミングを表す情報として捉え，食物の違いと咀嚼回数に基づいて疑似咀嚼音の振幅と持続時間を変化させるように疑似咀嚼音呈示装置を改良した.その結果，咀嚼回数に応じて疑似咀嚼音の持続時間と音量を調整できるようになり，改良前よりも自然な疑似咀嚼音を生成できるようになった.