Hello Shantanu:

If you want ALL of the area above the baseline (not just the area above
the baseline until the first baseline crossing), you could try the following:

First, perform the deconvolution analysis, with both a bucket dataset output
and the fit time series output:

3dDeconvolve \
-input myData+orig -polort 1 -num_stimts 1 \
-stim_file 1 myStim.1D -stim_label 1 myStim -stim_maxlag 1 10 \
-progress 1000 -fout -bucket myData.bucket \
-fitts myData.fitts

Now, use 3dcalc to subtract the baseline, and set negative values to zero:
3dcalc \
-a myData.fitts+orig -b myData.bucket+orig'[0]' -c myData.bucket+orig'[1]' \
-expr "70*(a-b-c*l)*step(a-b-c*l)" -fscale -prefix myData.detrend

Note that the above expression first multiplies by 70, which is the number of
time points - maxlag. Also, note that "l" in the above expression is NOT a "1".

Next, use 3dTstat to calculate the mean (which, since you have already
multiplied by the number of time points, the "mean" is actually "area"):
3dTstat -prefix myData.area myData.detrend+orig'[10..79]'

The output dataset, myData.area+orig, should contain the area in the fitted
response above the baseline.

Doug Ward