The sampling frequency and the corresponding five decomposition levels  corresponding to the eeg signal bands  are given in the link https://arxiv.org/ftp/arxiv/papers/1307/1307.7897.pdf.

Assuming the sampling frequency is fs, then the highest frequency that the signal could contain, from Nyquist’ theorem, would be fs/2. Frequency bands corresponding to five decomposition levels for wavelet db4 with sampling frequency

of 173.6 Hz of eeg signals were listed in Table . The signals were decomposed into details D1-D5 and one final approximation A5.

TABLE I

FREQUENCY BANDS CORRESPONDING TO DIFFERENT DECOMPOSITION

LEVELS Di,DECOMPOSED SIGNALS FREQUENCY BANDS(HZ) 

D1 43.4-86.8  (noises)

D2 21.7-43.4  (gama)

D3 10.8-21.7 (beta)

D4 5.40-10.8  (alpha)

D5 2.70-5.40 (theta)

A5 0.00-2.70 (delta)

So the sampling frequency  must be chosen to cover all possible frequency bands in the signal including the high frequency noise assumed in the band D1. Then comes the  gama, the beta, the alpha, the theta and the delta which is lowest frequency wavelets.

Best wishes

The same frequencies as used in Fourier-based spectrum analysis. Wavelets are NOT a good alternative way for performing spectral analysis.

The sampling frequency and the corresponding five decomposition levels  corresponding to the eeg signal bands  are given in the link https://arxiv.org/ftp/arxiv/papers/1307/1307.7897.pdf.

Assuming the sampling frequency is fs, then the highest frequency that the signal could contain, from Nyquist’ theorem, would be fs/2. Frequency bands corresponding to five decomposition levels for wavelet db4 with sampling frequency

of 173.6 Hz of eeg signals were listed in Table . The signals were decomposed into details D1-D5 and one final approximation A5.

TABLE I

FREQUENCY BANDS CORRESPONDING TO DIFFERENT DECOMPOSITION

LEVELS Di,DECOMPOSED SIGNALS FREQUENCY BANDS(HZ) 

D1 43.4-86.8  (noises)

D2 21.7-43.4  (gama)

D3 10.8-21.7 (beta)

D4 5.40-10.8  (alpha)

D5 2.70-5.40 (theta)

A5 0.00-2.70 (delta)

So the sampling frequency  must be chosen to cover all possible frequency bands in the signal including the high frequency noise assumed in the band D1. Then comes the  gama, the beta, the alpha, the theta and the delta which is lowest frequency wavelets.

Best wishes

The problem in not clear by query. If you are asking about sampling frequency to extract standard band such as Gama, Beta, Alpha....then Mr. Abdelhalim Zekry explanation will solve your problem. Otherwise you should know about sampling frequency of acquired data then follow the decomposition rule as --  first detail coefficients as  fs/4 to  fs/2, then second as  fs/8 to  fs/4.......... nth level as  fs/2n+1 to 0 depends on desired frequency bands.