After the multi-channel implant had been surgically inserted, the first studies on the patient showed the physiological limitations in coding sound reported above were also reflected in the perception of sound with electrical stimulation. Firstly, rate of stimulation could only be discriminated as pitch up to 300 pulses per second. This made it suitable for distinguishing voiced sounds such as the consonant /b/, but not /p/. Secondly, when the speech frequencies were filtered into a number of separate frequency bands (top), and the energy in those bands used to stimulate the electrodes on a frequency place coding basis, they were perceived as timbre rather than pitch. Timbre is the quality of a sound that distinguishes musical instruments playing the same notes. Furthermore, site of electrode stimulation could be distinguished on the basis of the sound being "sharp" or "dull."
Our first speech studies filtered the speech waveform (centre) with filters (top), but speech understanding was poor due to unpredictable variations in loudness due to the interaction of the electrical currents from each filter. This established the principle of not having simultaneous electrical stimulation for multiple electrodes. In summary, the perceptual studies demonstrated that the interface between an implant electrode and the brain was an electro-neural “bottle-neck” limiting the coding of acoustic signals for brain processing and perception (bottom).