Sound excerpts:

[01]
[02]
[03]


Dr. Sonja Gruen, Neuroinformatics, Freie Universität Berlin

The Neurophone

It was proposed by Donald Hebb (1949) that cortical activity is organized in functional groups ('cell assemblies') that form the building blocks of computational processes by the coordinated activity of the participating neurons. This hypothesis prompted neuroscientists to record from multiple neurons at a time, to observe their simultaneous activity. Such parallel activity is analyzed for correlated activity e.g. in terms of coincident spiking activity (e.g. Riehle et al, 1997; Singer, 1999) and/or spatio-temporal patterns (e.g. Prut et al, 1998), and results are typically displayed for visualization. Another approach would be to use our auditory perceptional abilities in order to identify temporal structures in the neuronal activity. The neurophone, introduced by Aertsen & Erb (1987) suggests the idea to 'audialize' neuronal activity by presenting their simultaneous activity in a form compared to music.

Electrical signals recorded from electrodes in the brain next to neurons are reduced to the time stamps when action potentials (spikes) occurred, since the amplitude of spikes are of unique height and do not carry information. Thus, doing this for multiple neurons recorded in parallel, we get for each neuron a time series of marked points in time, which we treat as the time of occurrences of musical notes. Each neuron we treat as a single 'voice' and assign a different tone to each of them. The result is a piece of 'neuromusic', in which each neuron contributes with its tone at each time it emitted a spike.

The pieces we are listening to (implemented by Sonja Gruen (http://www.mpih-frankfurt.mpg.de/global/np/staff/gruen.htm) and Michael Erb (http://www.medizin.uni-tuebingen.de/~mlerb/), 1992) were translated from recordings 1) from prefrontal cortex in awake behaving monkey involved in a sensorimotor behavioral task (Vaadia et al, 1989), 2) from data of neural network simulations of synfire chains (Diesmann et al, 1999) and 3) from data recorded in primary motor cortex of awake behaving monkey involved in a delayed-response handmovement task (Riehle et al, 1997).

References:
Aertsen A, Erb M (1987) Scherzo cerebri, ma non troppo: das Neurophon. (in German) In: Wahrnehmungen des Gehirns, pp 67-86. Aertsen A, Palm G, Popp M, Schüz A (eds). Tübingen (FRG): SynAPPS. http://www.brainworks.uni-freiburg.de/
Diesmann M, Gewaltig M-O and Aertsen Ad (1999) Conditions for stable propagation of synchronous spiking in cortical neural networks. Nature 402: 529-533 http://www.chaos.gwdg.de
Prut Y, Vaadia E, Bergman H, Haalman I, Slovin H and Abeles M (1998) Spatiotemporal Structure of Cortical Activity: Properties and Behavioral Relevance. J Neurophysiol 79(6): 2857--2874.
Riehle A, Gruen S, Diesmann M and Aertsen A (1997) Spike Synchronization and Rate Modulation Differentially Involved in Motor Cortical Function. Science, 278:1950-1953. http://lnf.cnrs-mrs.fr/crnc/cortex/
Singer W (1999) Neural synchrony: a versatile code for the definition of relations. Neuron 24: 49--65. http://www.mpih-frankfurt.mpg.de/global/np/staff/singer.htm
Vaadia E, Bergman H and Abeles M (1989) Neuronal Activities Related to Higher Brain Functions - Theoretical and Experimental Implications", IEEE Trans Biomed Eng 36(1): 25-35. http://www.md.huji.ac.il/depts/physiology