Artificial neural networks: new perspectives in neuropsychiatry

Background

In the last decades computer models are central to scientific disciplines.

Artificial neural networks are constructed according to fundamental Biology principles, and their properties are investigated, in order to provide theoretical background about the functioning of the human brain, and possible mechanisms underling mental illnesses.

The aim of this study is to investigate if there is an increased scientific interest in the study of artificial neural networks and to review their successful applications in neuropsychiatry.

Materials and methods

We conducted a search in the MEDLINE database from 1985 until 2005 using the following key words, grouped in 5-year periods: Artificial neural network; Artificial neural network brain.

The key-word "brain" was used in order to limit the search to neuropsychiatry related papers, whereas using the other 3 words alone, yielded results from all medical specialties.

Results

Table 1

Discussion

There is evidence for a continuously growing scientific interest in neural network models within the last two decades in medicine and especially in neuropsychiatry, as shown by the increasing number of relevant published papers. Most recent papers, in which neural network models give plausible mechanisms for cognitive functions such as associative memory, and for psychiatric symptoms such as delusions and auditory hallucinations, are reviewed.

The issue of how the brain represents information at the neuronal level is crucial because it sheds light from another perspective to cognition and mental illnesses and provides a potential target for therapeutic interventions.