In several studies about "Somatosensory evoked potentials (SEPs)", researchers investigate high-frequency oscillations (>400Hz). Besides the specification of the amplifiers (sampling rate and filtering setting), would it be meaningful to discuss/analyze these HFOs as it is recorded by "scalp EEG electrodes"? in most EEG studies, frequency beyond 100 Hz (even 50Hz) has been considered as noise.
Would you please share your thought on this?
Hello, i understand your point, SEP recordings show HFO even with scalp electrodes (given that salmpling rate is high enough); However recordings need to be done with attention and in particular you must avoid muscular contraction and AC artifacts, you have to average at least 1000/2000 ERP to see a proper response.
Several studies show how Hig frequency oscillations seen at cortical SEP are the direct reflection of the fast activity recorded with subdural electrodes in human ore mice. (see the papers on SEP-HFO by Ozaki, Kimura or Insola).
According to Ozaki further guarantee of genuine HFO is the inversion of the fast oscillating component between frontal Fz electrode and Cortico parietal (Cp) electrode. This happens only in the cortical HFO component underlying N20 response ,this is because cortical activity is a tangential dipole wih opposite phase in those two electrodes, while muscular activity or noise would not have phase inversion between Fz and Cp.
Concerning EEG i have personally red papers claiming to see high frequency activity on scalp EEG, however we must be carefull since one thing is recording HFO in a controlled condition as SEP (in which we know that there are such oscillations) another thing is finding fast oscillation in scalp EEG and beeing sure that they are brain signal. I think that only stereo EEG has that capability. In clinical EEG we can only hypothesize that fast oscillations are due to brain activity if they are supported by strong anamnestic data (i.e. the existence of focal cortical dysplasia)
Agreed with Jacopo. Very well said.
Physically in EEG or ERP it is possible. You are right you will want to check your amplifier and you would need a high sampling rate, at least 2000. And a long enough resting state recording. For ERP you would need thousands of samples, as noted by Jacopo. The interpretation of the EEG would be a challenge though and even very small artifacts become trouble when looking at HFO
As far as I know, scalp behaves as a low-pass filter, entailing a bit problematic high-frequencies signal analysis from scalp EEG data. Instead, one uses intracranial EEG signals .
Article Crone NE, Sinai A, Korzeniewska A. High-frequency gamma osci...
- Badges
- Science topic