Pathological cerebral synchronization may severely perturb brain function as
observed in several neurological diseases like Parkinson's disease and
essential tremor. In patients that do no longer respond well to drug therapy,
depth electrodes are chronically implanted in target areas located in the
thalamus or the basal ganglia. To suppress the pathologically synchronized
firing, a permanent high-frequency (
100 Hz) stimulation is performed
(Benabid et al. 1991, Blond et al. 1992).
Although the therapeutic effects are impressive, there are nevertheless
significant drawbacks:
(i) The energy consumption of the permanent stimulation is quite high. Thus,
the generator (plus battery) has to be exchanged after 1-3 years by means of
an operation. (ii) Even more important is the fact that the permanent
high-frequency input is an unphysiological type of stimulation which causes
the stimulated target areas to adapt. In a number of patients the
amplitude of the stimulation has to be increased over the years, in order
to maintain the tremor suppressive effect. With increasing stimulation
amplitude, however, the probability of the occurence of severe side effects
(like dysarthria, dysaesthesia, cerebellar ataxia, psychotic symptoms)
increases.
With methods from synergetics (Haken 1983) and statistical physics (Kuramoto 1984) the concept of phase resetting (Winfree 1984) was extended to populations of interacting oscillators subjected to random forces (Tass 1999). This stochastic phase resetting approach has lead to the development of demand-controlled deep brain stimulation techniques (Tass 2001a-2001c, 2002a, 2002b). The latter work in a completely different way compared to the standard high-frequency stimulation: While the standard technique probably simply suppresses the neuronal firing in the target area (Wielepp et al. 2001), the novel techniques only desynchronize the firing whenever it gets pathologically synchronized. In this way, the novel methods intent to bring the neurons' dynamics as close to the physiological state (i.e. to the uncorrelated firing) as possible. The talk is about both theory and first experimental results.
References
Benabid AL et al. 1991 The Lancet 337 403
Blond S et al. 1992 J. Neurosurg. 77 62
Haken H Advanced Synergetics (Springer, Berlin, 1983)
Kuramoto Y Chemical Oscillations, Waves, and Turbulence
(Springer, Berlin, 1984)
Tass PA Phase Resetting in Medicine and Biology -
Stochastic Modelling and Data Analysis (Springer, Berlin, 1999)
Tass PA 2001a Europhys. Lett. 53 15
Tass PA 2001b Biol. Cybern. 85 343
Tass PA 2001c Europhys. Lett. 55 171
Tass PA 2002a Europhys. Lett. 57 164
Tass PA 2002b Biol. Cybern. (in press)
Wielepp JP et al. 2001 Clin. Neurol. Neurosurg. 103 228
Winfree AT The Geometry of Biological Time (Springer, Berlin, 1980)
Cerebral synchronization processes play an essential role under both physiological (Freeman 1975) and pathological (Llinás and Jahnsen 1982, Bergman et al. 1994) conditions. To detect and localize phase synchronization and stochastic phase resetting dynamics in the human brain non-invasively with magnetoencephalography novel methods have been developed:
1. Synchronization tomography (Tass et al. 2002):
First, the cerebral current source density is reconstructed in each
cerebral voxel (i.e. volume element) for each time 
by means of the
magnetic field tomography (MFT) (Ioannides et al. 1990). Next, the
phase synchronization analysis (Tass et al. 1998) is applied to each
voxel and to external reference signals such as muscular activity. In
this way brain/brain- and brain/muscle phase synchronization are
determined. It turns out that phase synchronization is a fundamental
coordination principle in cerebral motor control (Tass et al. 2002).
2. Phase resetting tomography: The cerebral current source density is reconstructed with MFT. Next, a stochastic phase resetting analysis (Tass 1999, 2002a, 2002b) is applied to each voxel as well as to all pairs of voxels and external signals. This enables the detection of transient stimulus-locked response clustering and transient stimulus-locked synchronization and desynchronization. In contrast, standard techniques like cross-trial averaging or cross-trial cross-correlation are not able to detect such processes and may even produce artifacts.
The talk is about the theoretical background of the novel methods, their application to experimental data, and their diagnostic relevance.
References
Bergman H et al. 1994 J. Neurophysiol. 72 507
Freeman WJ Mass action in the nervous system
(Academic Press, New York, 1975)
Ioannides AA et al. 1990 Inverse Problems 6 523
Llinás R and Jahnsen H 1982 Nature 297 406
Tass PA Phase Resetting in Medicine and Biology -
Stochastic Modelling and Data Analysis (Springer, Berlin, 1999)
Tass P et al. 1998 Phys. Rev. Lett. 81 3291
Tass PA 2002a Europhys. Lett. (in press)
Tass PA 2002b Chaos (in press)
Tass PA et al. 2002 (submitted)