Abstracts

RATIONALE: Animal studies, as well as invasive recordings from humans, have demonstrated prominent, very slow activity patterns (i.e. DC potential shifts) during eg. epileptic activity, ischemia, and sleep. Due to previous methodological difficulties there are, however, only few published non-invasive DC-EEG recordings on human patients with epilepsy or with any other clinical condition. We have recently solved these issues and developed a DC-EEG technique that is capable of long-term, bedside recordings. In this paper we describe practical details of the DC-EEG technique used at our center, and we discuss the technical issues that are requisite for a reliable recordings from human scalp.
METHODS: We use a custom made DC-EEG amplifier, and sintered Ag/AgCl electrodes, which are individually attached to skin (with collodion) by using custom tailored plastic holders. Software for data acquisition and analysis were written under Labview. We examined the DC recording properties of several commonly used electrode materials (Ag/AgCl, silver, gold, tin, stainless steel, and platinum). We examined the characteristics, as well as the potential means of elimination of several biological artefacts: skin potential, eye rotations (gaze shifts), tongue movements, respiration, and changes in intracranial pressure by Valsalva maneuver. Distribution of electric fields generated by eye and tongue movements were also evaluated with a 256 channel EEG device (Geodesic Inc., OR).
RESULTS: We show that our DC-EEG setup is capable of long term (at least 24 hrs) stable recordings. We discuss the specific requirements of the amplifier, electrodes, and software. Comparison of different commercial electrode materials demonstrates that only Ag/AgCl gives a stable and reliable long-term recording. Skin potentials are easily short-circuited (ie.eliminated) by scratching the skin, otherwise these long-term recordings would be severily contaminated with these millivolt scale responses and baseline drifts. Tongue movements produce slow (up to many seconds) potentials effectively picked up by mastoid electrodes, and with amplitudes comparable to those of very slow evoked responses (eg. CNV). Tongue artefacts can be minimized by using termporal or vertex referencing. Eye rotations also cause electric potentials recordable over the whole head. Use of appropriate montages and eye channels makes it possible to exclude these artefacts. Respiration, by changing the end-tidal CO2, causes a very prominent DC potential shift with amplitudes up to 2mV at vertex referred to mastoid. Finally, changes in intracranial pressure conditions by Valsalva maneuver results in a clear potential gradient between vertex and temporal/mastoid derivations, which implies that respiratory activity has to be carefully monitored during the recording periods of interest.
CONCLUSIONS: Long-term bedside recordings are technically rather easy with very achievable modifications to routine clinical EEG methods. Presence of several non-cortical sources of slow electric potentials requires attention. Our experience has shown, however, that with the precautions outlined in our paper reliable DC-EEG recordings may be readily performed providing a new insight into the mechanisms of epilepsy and other brain conditions (see abstracts by Thompson et al and Miller et al, this meeting).
[Supported by: Finnish Academy, University of Washington Regional Epilepsy Center]