The six papers in this special section focus on the most recent applications of computational intelligent systems in medicine. Some papers accepted for this issue (10 in total) were published earlier by mistake.

A model of the subthalamo-pallidal loop using integrate-and-fire neurons in conjunction with first order models of synaptic afferents, calcium and potassium currents, and a leaky sodium current exhibits activity similar to that recorded in the Subthalamic nucleus of Parkinson patients undergoing stereotactic neurosurgery and deep brain stimulation (DBS). In particular, by incorporating a negative feedback loop, the model produces inhibitory responses to individual DBS pulses and exhibits oscillations in the alpha and beta range (8-29 Hz). Moreover, an increase in either DBS stimulus energy or GPe spontaneous activity causes the oscillatory behavior to subside.

The primary focus of this paper was to develop a high-performance computer system for optimizing auditory stimuli based on neuronal feedback. Using the Algorithm Of Pattern EXtraction (ALOPEX) extra-cellular action potentials (APs) recorded from frog (Rana Pipiens) auditory neurons were used as feedback to optimize sound stimuli. This computer-based system works in real time to iteratively find the neuron's best excitatory frequency (BEF). Three programmable (positive and negative) threshold logic levels are used to collect 300 APs in response to normalized pure tones. Fuzzy logic is then used to separate up to five fuzzy centers (templates) from the 300 APs. The fuzzy centers are used for on-line fuzzy mapping of future responses. The five fuzzy centers allow the system to monitor up to five neighboring neurons. To study the auditory neurons of the frog, one, two, and three simultaneous tones are used as the stimulus for optimization of the best combination of frequencies. Testing with the response calculated as a parabolic function of a single best frequency demonstrated system dynamics and reliability for up to nine simultaneous tones. Experiments using one pure tone and ten stimulus presentations per iteration showed that the automated system is able to repeatedly converge to the best frequency within 100 iterations. Studies using one, two, and then three pure tones played simultaneously on the same group of neurons has shown that these tones converged on the same best frequencies by properly mixing the tones available to produce the optimal complex sound.

The main purpose of this work is to create a physiologically realistic computational model of the retina by providing a flexible, real-valued three-dimensional architecture. This model exhibits the overall spatio-temporal response of the network as a whole, as well as the temporal responses of individual cells in the different layers of the retina. The network performs both functions in a consistent way. The model can be expanded to include color vision, Mach band effects. It can also be used to model other physiological systems and functions. The platform on which this neural network is built is Windows 98 using an object oriented approach.

This paper describes the application of a novel unsupervised pattern recognition system to the classification of the Visual Evoked Potentials (VEP's) of normal and multiple sclerosis (MS) patients. The method combines a traditional statistical feature extractor with a fuzzy clustering method, all implemented in a parallel neural network architecture. The optimization routine, ALOPEX, is used to train the network while decreasing the likelihood of local solutions. The unsupervised system includes a feature extraction and clustering module, trained by the optimization routine ALOPEX. Through maximization of the output variance of each node, and an architecture which excludes redundancy, the feature extraction network retains the most significant Karhunen-Loeve expansion vectors. The clustering module uses a modification to the Fuzzy c-Means (FCM) clustering algorithms, where ALOPEX adjusts a set of cluster centers to minimize an objective error function. The result combines the power of the FCM algorithms with the advantage of a more global solution from ALOPEX. The new pattern recognition system is used to cluster the VEP's of 13 normal and 12 MS subjects. The classification with this technique can, without supervision, separate the patient population into two groups which largely correspond to the MS and control subject groups. A suitable threshold can be chosen so that the recognizer chooses no false negatives. The use of multiple stimulation patterns appears to improve the reliability of the decision. The reasoning of most neural networks in their decision making cannot easily be extracted upon the completion of training. However, due to the linearity of the network nodes, the cluster prototypes of this unsupervised system can be reconstructed to illustrate the reasoning of the system. In this application, this analysis hints at the usefulness of previously unused portions of the VEP in detecting MS. It also indicates a possible use of the system as a training aide.

Using conventional very large scale integration (VLSI) technology, a flexible and comprehensive neuromime circuit has been implemented in silicon for the purpose of modeling nerve networks from living organisms. Based on the "integrate-and-fire" model of neuronal threshold, the circuit was fabricated in two-micron CMOS with double-level metal. It occupies 0.6 square millimeters of die area, and requires only a few passive biasing components off-chip. The neuromime circuit offers many continuously variable parameters, including excitatory and inhibitory sensitivity and persistence, refractory duration and strength, and overall speed of operation. The circuit offers free and continuous access to waveforms for presynaptic membrane potential, postsynaptic membrane potential, and threshold potential. As such, it is amenable to many secondary behavioral characteristics, such as postinhibitory rebound, fatigue, facilitation, and accommodation. Being power-efficient, compact, and noise immune, it is ideal for assembly into networks and interfacing to biological counterparts.

The commenters refer to the above mentioned paper by Wolpert and Micheli-Tzanakou (ibid. p.955-61, 1993), and point out the oversight of previous activity in circuit design. In their opinion the paper in question does not seem to represent a novel approach either in circuitry or computationally. In reply, the authors acknowledge their oversight of the conference papers cited by the commenters, but point out that they never suggested that a simple shunting lateral inhibitory (SLI) pair was a novel circuit, and the overall objective of their effort was simply to carry analysis of SLI to a higher dimension.

The neurological process known as lateral inhibition (LI) has long been acknowledged as a critical operation for the preprocessing many types of sensory stimuli. In the mammalian retina, LI is utilized to enhance visual images by performing differential amplification on the pixels from which the image is composed. In this study, LI is implemented using VLSI-based models. These models consist of small two-dimensional arrays of generalized sensory pixels, each of which inhibits, and in turn is inhibited by, each of its immediate neighbors. Two custom CMOS array prototypes circuits have been designed, fabricated, and characterized. Test results indicate that both circuits are able to impart contrast to arbitrary two-dimensional geometric images in a flexible yet stable manner, and do so immediately and simultaneously. These arrays thus offer a level of performance not attainable by software methods, making this method well suited for machine vision systems that utilize parallel architectures.

The September 1990 special use on neural networks of the IEEE Engineering in Medici and Biology Magazine is described. This issue contains of papers dealing with various aspects and applications of neural networks in a wide spectrum of biomedical engineering problems.

The author examines the physiological aspects of the visual systems of animals and humans and several models that explain or try to explain physiological or psychophysical data. Particular attention is given to the decomposition of patterns by the human brain. Both digital and analog implementations are discussed and some projections to future applications are also mentioned. The development of a neuronal computer vision system is considered.>