The present protocol describes the transcranial application of oscillating current stimulation (0.75 Hz) for inducing slowly oscillating potential fields in underlying neocortical tissue during early nocturnal non-rapid eye movement (NonREM) sleep. In itself the application of weak currents to the human brain via scalp electrodes has been employed, in the course of history, in numerous variations, i.e., using different current intensities, current waveforms, stimulation durations and different positions and sizes of stimulating electrodes. However, currents were applied mainly during wakefulness and frequently the attempts served therapeutic purposes. The aim of the oscillating current stimulation is to boost endogenous slow oscillatory electroencephalographic (EEG) activity during sleep, taking advantage of resonance effects occurring in neuronal networks for defined oscillatory rhythms. Aside from transcranial stimulation the protocol requires polysomnograhic recordings.

The idea of applying constant and oscillating currents to the human scalp and brain dates back to way before the 19^th century. At the beginning of the 20^th century electric stimulation, often together with psychopharmacological treatments, was employed in an attempt to aid various psychiatric conditions and also insomnia. However, results of different laboratories were often not consistent, due to the diversity in stimulation parameters which mostly also lacked a detailed documentation. As the reliability of psychopharmacological treatments developed, the interest in transcranial stimulation with electrical currents for clinical use in humans faded. Within the last 60 years, methodologically well documented systematic studies have lent support for the efficacy of transcranial current stimulation to affect neuronal excitability within the brain ^1-6. Some of these studies in humans showed that transcranial direct current stimulation changes excitability within the visual and motor cortex ^7,8 as well as behavioural indicators of cortical processing such as reaction time and working memory performance ^9-13.

Transcranial stimulation with currents is different from the transcranial magnetic stimulation (TMS) of the brain which recently has become a widely used technique in human neurosciences and clinical neurology to excite or inhibit local regions of the neocortex ^14,15. Different from TMS which is typically applied as single or repetitive pulses, transcranial stimulation with currents aims to induce slow and gradual changes in potential fields in underlying brain tissue. Also due to the resistance and capacitor effects of the scalp, skull, cerebroventricular fluid etc. transcranial stimulation with currents reaches more widespread areas of the cortical surface.

The parameters for the present protocol are based on the concept that the oscillating current stimulation induces potential fields in underlying cortical tissue which are similar in size and frequency to those naturally occurring in these networks during defined states of sleep and wakefulness. It is assumed that the oscillating potential fields induced by stimulation through mechanisms of resonance can potentiate endogenous rhythmic EEG activity present in these tissues. Neurons are indeed capable of synchronizing to weak oscillating fields as small as 295 µV mm^-1 (140 µV mm^-1 rms; ref. 16). In vitro experiments indicated that the synchronizing effect is even stronger for neuronal networks than for single neurons and persists in the absence of functioning chemical synapses ¹⁷. In line with this concept it should be possible to boost other endogenous EEG rhythms by applying corresponding stimulation frequencies, (e.g., during a state of wakefulness) at a time when neocortical networks are particularly sensitive to tune into the respective rhythm.

Our experiments focused on the < 1 Hz slow oscillation during sleep which groups delta and spindle EEG activity in humans and is presumed to contribute to memory consolidation during sleep ^18-20. Thus, the present protocol of slow oscillation stimulation aims to enhance slow rhythmic oscillations and associated spindle activity when applied during early nocturnal NonREM sleep which is a time of emergent endogenous slow oscillation activity. It is, hence, necessary to monitor also sleep and sleep stages, to set the beginning of stimulation. Sleep stages are conventionally determined by the criteria of Rechtschaffen and Kales ²¹. Slow oscillation stimulation begins shortly after the subject has attained stable NonREM sleep for the first time after sleep onset, i.e., specifically after online scoring of polysomnographic recordings confirms the presence of eight consecutive 30-s epochs of sleep stage 2 or a deeper NonREM sleep stage.

Aside from monitoring endogenous brain electric activity for determining sleep stage the investigator might want to record EEG from other scalp locations to identify immediate effects of the stimulation. Note, however, that intervals during acute stimulation cannot necessarily be scored or analyzed for cortical activity due to excessive signal artifacts. Therefore, we stimulated repeatedly for 5-min intervals and introduced 1-min stimulation-free intervals between these periods of slow oscillation stimulation which allows analyzing immediate after-effects of the stimulation.

In the described experiments we used a stimulator (Electronics Department, University of Lübeck) which produces a regulated current oscillating at a frequency of 0.75 Hz between a peak positive value of a predetermined strength of 260 µA and zero current. Strength of slow oscillation stimulation is derived from estimated potential fields induced in underlying neocortical tissue that should not exceed the size of potential fields naturally occurring in these tissues ^22-24. Oscillating stimulator output can be anodal or cathodal and it is debatable whether the polarizing component may modify effects. We applied current via electrodes attached bilaterally over left and right prefrontal cortex (F3, F4) and to both mastoids. Positioning of electrodes was chosen based on the predominant generation of slow oscillations in the prefrontal cortex ²⁵. Cortical effects of transcranial electrical stimulation depend strongly on the positioning of stimulation electrodes ²⁶. The positioning and possibly also the number of stimulation electrodes should be adapted for oscillating current stimulation which is to correspond to another rhythm.

It is very important to take into account, that current strength, size of stimulation electrodes as well as the stimulation frequency and total duration are all relevant for the safety of the stimulation. Also consider in placing electrodes that the current density in the cortex immediately underneath a cranial suture may be several times higher than under the non-suture area ²⁴. It is to note, that the electrical parameters used for oscillating current stimulation are closely comparable to those used normally with transcranial direct current stimulation. However, they are not at all comparable with the high intensities used in some studies with pulsed transcranial electric stimulation or with electro convulsive therapy. For detailed discussions on resulting intracerebral currents and safety limits see also refs. 8,23,27-30.

To capture immediate effects of current stimulation on EEG activity, recording at additional sites is required, as noted above. Slow oscillation stimulation during early nocturnal NonREM sleep modifies endogenous EEG activity and this is accompanied by a functional effect on concurrent central nervous processing, i.e., on processes of memory consolidation ³¹. Anticipated results are seen in the mean response of a subject sample, with each subject undergoing stimulation and a sham stimulation control condition where no current is applied. Although oscillating current stimulation may be adapted for use with various parameters as discussed above, specifications given below are for the slow oscillation stimulation as employed by Marshall et al.³¹