Is it necessary to reconstruct the signal after Single-level discrete 1-D wavelet transform (dwt)

hi friends, i have to seperate the alpha band a channel (AF3)of eeg signal. Sampling frequency is 128 Hz. i applied dwt function like this:

data=xlsread('E:\color\neha.xlsx'); % location of data file data= data'; value1=data(1,:); chan1= value1-mean(value1); % normalisation [cA,cD] = dwt(chan1,'db4'); [cA1,cD1] = dwt(cA,'db4'); [cA2,cD2] = dwt(cA1,'db4'); [cA3,cD3] = dwt(cA2,'db4'); [cA4,cD4] = dwt(cD3,'db4');

now as per my thinking cD3 will be the signal which ranges 8-12 hz ie. alpha band. Can any one clarify if i am correct? Or is here any need of wavelet reconstruction command to get the alpha band.