NICI, Department of Psychology, University of Nijmegen, Nijmegen, The Netherlands
The cortical EEG of rats contains transient phenomena such as sleep spindles and spindles belonging to the intermediate state. Various in and outbred strains of rats show, besides these transients, spontaneously occurring spike-wave discharges (SWD) in the EEG (van Luijtelaar & Coenen, 1986). During these SWDs the ability to respond to external stimuli is diminished, but not completely absent. This may suggest that sensory processes are disturbed during SWDs. The SWD represents thalamocortical oscillations: by studying these, information can be obtained about the functioning of the thalamus and its cortical connections.
Electrophysiological studies, however, require often long-time registration and thus the task of detection of these transient aberrant phenomena becomes tedious. Furthermore, in order to investigate sensory processing during SWDs it is imperative to stimulate during these phenomena. This is difficult because the phenomena are short in duration (± 5 to 6 s) and the time of stimulus presentation after onset has to be kept constant, since frequency modulations occurs (Drinkenburg et al., 1993). In order to solve these problems a system is developed that is capable of detecting SWDs both on- and off-line.
The automatic detection is performed by digitising the EEG at 200 Hz and subsequently calculating the absolute difference of two consecutive digitised samples. This value is called the steepness of the EEG. On the basis of this steepness, the maximum steepness over a time period of 0.25 s is derived. If the average value over a time period of 0.25 s exceeds a certain level for 1 s, an aberrant transient is detected. This level is automatically estimated on the basis of the EEG during wakefulness.
This system is evaluated by comparing its on-line detection of aberrant EEG phenomena of 9 rats (2 h each), to the consensus of two experts. The consensus detected in total 405 SWDs. The automatic system detected 392 phenomena correctly, 97%. Thirty-six incorrect detections were made, mainly consisting of movement artefacts.
This performance of the system is remarkable, considering the high morpho-logical variation between and within SWDs. A second merit of this system is the relative simplicity of the algorithm. This simplicity makes on-line, as well as high speed off-line detection possible, using an relative inexpensive personal computer. So it seems that this system possesses enough validity to be a valuable and easy to use instrument in experiments.
References
Drinkenburg, W.H.I.M.; van Luijtelaar, E.L.J.M.; van Schaijk, W.J.; Coenen, A.M.L. (1993). Physiology & Behavior 54:779-783.
Luijtelaar, E.L.J.M. van; Coenen, A.M.L. (1986). Neuroscience Letters 170:393-397.