A Bioassay System Using Bioelectric Signals from Small Fish

Abstract

Although the quality of tap water is generally examined using chemical assay, this method cannot be used for examination in real time. Against such a background, the technique of fish bioassay has attracted attention as an approach that enables constant monitoring of aquatic contamination. The respiratory rhythms of fish are considered an efficient indicator for the ongoing assessment of water quality, since they are sensitive to chemicals and can be indirectly measured from bioelectric signals generated by breathing. In order to judge aquatic contamination accurately, it is necessary to measure bioelectric signals from fish swimming freely as well as to stably discriminate measured signals, which vary between individuals. However, no bioassay system meeting the above requirements has yet been established. This paper proposes a bioassay system using bioelectric signals generated from small fish in free-swimming conditions. The system records signals using multiple electrodes to cover the extensive measurement range required in a free-swimming environment, and automatically discriminates changes in water quality from signal frequency components. This discrimination is achieved through an ensemble classification method using probability neural networks to solve the problem of differences between individual fish. The paper also reports on the results of related validation experiments, which showed that the proposed system was able to stably discriminate between water conditions before and after bleach exposure.