Tractography, Tolerance and number of tracts
Maryam Shapourjani
Hello Dear Frank,
I hope you are doing well.
My work focuses on dMRI association pathways in term-born neonates between 37 and 44 weeks of age.
- In the tractography batch step of the software, three default values are entered for tolerance: 22, 24, and 30.
a. Is it necessary to enter three values in this section, or is it also possible to enter only one value?
b. Based on your answer to question (a), for tractography of neonatal data focused on analyzing neonatal shape metrics and modeling the development of these features with age, would you recommend using a single tolerance value or three values?
- Regarding the number of tracts extracted for each pathway, you mentioned that at least 2,000 tracts should be extracted for each pathway in order to conclude that the number of tracts is sufficient to cover the region of interest.
a. Does this also apply to neonatal data, considering differences in tract development between 37 and 44 weeks of age? For example, should at least 2,000 tracts be extracted for both a 37-week-old neonate and a 44-week-old neonate?
b. Does this criterion apply to all association pathways, or are there pathways for which a different minimum number of streamlines would be more appropriate?
Thank you very much for your time and guidance.
Best regards,
Maryam Shapourjani
Frank Yeh
Hi Maryam,
Thank you for the detailed questions.
For AutoTrack tolerance, you can enter a single value. Multiple values are mainly useful as a parameter sweep or retry strategy, where a higher tolerance allows more deviation from the atlas tract pattern and may recover difficult tracts, but it may also include more false-positive streamlines.
For neonatal data, I would be cautious because the built-in recognition atlas was developed mainly from adult data. If your goal is quantitative modeling of development from 37 to 44 weeks, I would recommend using a consistent tolerance setting across all subjects, or using a clearly defined rule for choosing among multiple tolerance values. Otherwise, differences in the selected tolerance may become a confound in the shape metrics.
Regarding the number of tracts, the 2,000-streamline suggestion is a practical guideline to make sure the pathway is reasonably sampled. It is not a biological cutoff, and it may not apply equally to every tract or every neonatal scan. Small pathways, immature pathways, or lower-quality scans may not always reach this number. In those cases, I would rely on visual QC and test whether the tract shape metrics are stable when you increase the number of seeds or maximum tracts.
For comparison across age, I would avoid interpreting the raw streamline count itself as a developmental measure, because streamline count is affected by tracking parameters, image quality, and tract size. It is better to use the same tracking settings for all subjects and then analyze more stable tractometry or shape metrics after QC.
Best,
Frank
On Wed, 10 Jun 2026 23:45:46 -0700 (PDT), Maryam Shapourjani shapourja...@gmail.com wrote:
Hello Dear Frank,
I hope you are doing well.
My work focuses on dMRI association pathways in term-born neonates
between 37 and 44 weeks of age.
- In the tractography batch step of
the software, three default values are entered for tolerance: 22, 24, and
a. Is it necessary to enter three values in this section, or is it also
possible to enter only one value?
b. Based on your answer to question (a), for tractography of neonatal
data focused on analyzing neonatal shape metrics and modeling the development
of these features with age, would you recommend using a single tolerance value
or three values?
- Regarding the number of tracts
extracted for each pathway, you mentioned that at least 2,000 tracts
should be extracted for each pathway in order to conclude that the number
of tracts is sufficient to cover the region of interest.a. Does this also apply to neonatal data, considering differences in
tract development between 37 and 44 weeks of age? For example, should at least
2,000 tracts be extracted for both a 37-week-old neonate and a 44-week-old
neonate?b. Does this criterion apply to all association pathways,
or are there pathways for which a different minimum number of streamlines would
be more appropriate?Thank you very much for your time and guidance.
Best regards,
Maryam Shapourjani
–
You received this message because you are subscribed to the Google Groups “DSI Studio” group.
To unsubscribe from this group and stop receiving emails from it, send an email to dsi-studio+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/2594508d-31ac-416f-831b-ed8aeab80ef3n%40googlegroups.com.