Fiber Orientation during Tractography

[Christian Gewinn (Christent)]

Hi Frank,

After examining the data we extracted during automated tractography using the along tract algorithm you imlimented end of last year, we got surprised by the results, me and my coworkers have been trying to confirm the fiber orientation to make sure what we report is correct but havent been able to do so confidently. As such I would like to ask how the along tract data is extracted and sorted "under the hood". 

Are fiber data sorted segment wise, using the starting point found through the definition of the left/right setting during the call or are all datapoints just sorted from left to right/right to left globally. Sorry for reposting but this is regretably very important for our work.

Cheers,

Christian Gewinn

[Frank Yeh]

Hi Christian,

Sorry for the long delay in getting back to you. I needed some time to go through the code carefully.

In DSI Studio, the along-tract direction is determined as follows. At the end of fiber tracking, for each trajectory, the head and tail positions are used to compute a head-to-tail vector ((v_x, v_y, v_z)). DSI Studio then takes the absolute values ((|v_x|, |v_y|, |v_z|)) to identify the dominant dimension—that is, the dimension with the largest absolute value. If the head-to-tail vector component along this dominant dimension is negative, the trajectory is reversed.

Hope this clarifies things.

Best regards,
Frank