Modelling stabilograms with hidden Markov models.

Hidden Markov models are an effective computational method for modelling and interpreting digital signals of biological, as well as other, origin. In the current investigation, we explored whether hidden Markov models can be used to control and represent phenomena in human balance signals recorded from subjects standing on a force platform. Additionally, our aim was to classify healthy controls and patients who suffered from Meniere's disease into their own classes. Hidden Markov models were capable of these tasks and of overcoming such disturbances as noise and other unforeseen perturbations in balance signals, which are inherently complex and possibly difficult to visually specify.

Towards virtual reality stimulation in force platform posturography.

We developed a stimulation technique on the basis of virtual reality methods for balance investigation performed in balance laboratories of otorhinolaryngological clinics and institutes of occupational health. Such a stimulation technique is greatly progressive in the sense that by creating virtual moving views and "virtual worlds" inside which the subject is located it is possible to make effective stimuli that would be very difficult or even impossible to set in any real environment. We tested our system on healthy subjects and found out that this kind of virtual reality stimulation system is very useful for balance analysis.

Postural control as assessed with virtual reality.

We studied the use of virtual reality technology as a stimulus in balance examinations. A pilot study was made using a small group of healthy subjects to investigate the effect of alcohol and virtual reality stimulus on the subjects' balance. The tests showed that blood alcohol concentration accounted for almost 50% of the increased lateral body sway velocity. The new stimulus technique based on virtual reality technology seems to be effective and flexible for postural investigations.

Lossless compression of emission tomography images.

An adaptive lossless compression method using predictive coding with the Rice codes is proposed for low resolution three-dimensional biomedical images. The method is evaluated with the brain positron emission tomography (PET) images. The results are considerably better than with static arithmetic coding.

Lossy compression of auditory brainstem response signals.

Compression of digital signals is frequently utilised in numerous applications like music and speech signals. So far compression of biomedical signals has concentrated almost purely on electrocardiography. We recently opened compression studies on biomedical signals concerning otological measurements performed in accord with hearing and balance problems of humans. In this paper we shall present results for auditory brainstem response signals gained with traditional and modern compression methods as well as one of ours.

Lossless compression of eye movement signals.

Compression of digital signals is widely employed in several areas such as speech signals. Concerning biomedical signals almost merely ECG signals have so far been compressed, in fact, in very many applications. To our knowledge, this compression study is the first one presented in the case of eye movement signal which are investigated in several medical specialties. In this study eye movements are electro-oculographically recorded. Tests performed to signals of patients of an equilibrium laboratory gave compression ratios generated with essential lossless compression methods.