What could be heard and EDF as a sound format

Possibly we could identify some situations in which the use of audification could be useful in other physiological signals

• If we want to detect something with a frequency signature different from the background activity. Epileptic seizures are characteristically associated with brain chirps.
• If we want to detect changes in the frequency of the signal background. An example are sleep recordings but other long recordings such as surgical monitoring or Intensive Care Unit recordings also seem appropriate.
• If we want to relate two signals we could use stereo files. As an example we could follow the interhemispheric synchrony of right-left hemispheric activity.
• If we want to get some insight into the temporal appearance of a transient. Hearing is very appropriate to define the rhythm of a transient. As an example, in EMG end-plate potentials (random discharge) and Motor Unit Potentials (rhythmical discharge) can be identified even if the shape of the transient is not available.
• If we want to get some insight into the stable vs. intermittent nature of the frequencies present. In this sense we could use audificaton to evaluate the so-called microstructure of sleep (link to a PDF file).

Of course we can design tools adapted to address these goals. The advantage of audification is that we gain some intuitive insight different from the insight obtained with other methods. Many terms in EEG and sleep recording (lambda waves, spikes, sleep spindle, saw-tooth waves and others) denote the visual analogy of the context where they were first described. If sleep elements would have been heard instead of seen probably they would have other names.

The similarity between EDF(+) or other formats used in Clinical Neurophysiology and other sound formats is remarkable. As a matter of fact, EDF+ could also be used to store sounds. The following would be some ot the properties of EDF as a sound format

• Different sampling would allow adaptation to any circumstance. Imagine an alien thinking in MHz and a human speaking until 20 kHz. You can adapt each channel to record the telepathic conversation without waste of resources (without entering in the extraterrestrial intelligence you can imagine many other situations such as communicating with different bands, where different sampling rate would be convenient)
• Annotations allow a transcription (e.g., to include the transcription of the sound or the score of the piece) in the same file that stores the sound
• Amplitude is very precisely calibrated and you can make the audition very precise
• Discontinuous signals can be encoded, e.g., to store periods of silence with no coding at all
• A lot of segments can be stored in the same file (e.g., to store the catalog of the pronunciation of the different phonemes by a person).

Coming back to the proposed approach (an adapted program that allows the direct audition of the great amounts of physiological signals in EDF recordings), other alternatives such as using a mathematical program (Matlab, R, Octave, Scilab or others) to create sounds or having a plugin allowing the audition of physiological signals connected to the viewer are possible but they are more difficult to use or are not currently available.