A constant-voltage and constant-current stimulus isolation unit.

When action potentials from a stimulated biological preparation are recorded i t is essential to isolate the stimulus resistively and capacitively from the ground in order to minimize stimulation artifacts. Such isolation can be achieved by using a transformer of special low-capacitance construction in the output of a stimulus generator (Donaldson 1958). The most widely used arrangement for the isolation, mentioned previously , is the Schmitt-type isolation unit (Schmitt 1948). It consists of a driven radio-frequency oscillator with its output rf link coupled to the preparation. The most recent isolation units are based on the light transmission of electrical stimulus from the generator to the preparation (Dawson et al. 1960, Ross and Bassett 1967). By using the stimulus isolation unit in the output of a stimulus generator, the stimulating current flows entirely between stimulating electrodes, whether the preparation is grounded or not. When, however, electrodes made of thin wire (usually of 0.1 or 0.2 mm in diameter) are used, in this case for intracerebral stimulation of laboratory animals, the stimulus yields to appreciable distortion. When the constant-voltage type stimulus isolation unit (device with low output impedance) is used a current pulse yields to distortion. It is caused by the electrochemical polarization of the elect?odes used for stimulation (Weinman and Mahler 1964, Greatbatch 1967). The small contact areas of the electrodes with the preparation are the cause of high current densities at the metal-preparation interface, resulting in electric polarization with the electrode behaving as a capacitance. The typical distortion of a rectangular current pulse is shown in Fig. l b. This distortion makes the measurement of the stimulating current difficult. Since it is generally accepted that