fix: Ignore non-steady-state volumes in T2*/S0 estimation

[tsalo]

Closes #3507. As of 25.1.0, tedana has an --exclude parameter in its t2smap CLI, which allows users to ignore certain volumes when estimating T2* and S0. This PR bumps the tedana version and leverages that parameter to ignore dummy scans. If it proves useful, we could also ignore high-motion volumes in the future.

Changes proposed in this pull request

• Increase minimum tedana version from 25.0.0 to 25.1.0.
• Use estimated number of dummy scans to set exclude parameter in t2smap call, so that those volumes do not contribute to the estimation of T2*, which is in turn used to optimally combine the data across echoes. I doubt the dummy scans actually have much effect on T2* estimation, but it still stands to reason that they are not helping.

Documentation that should be reviewed

[tsalo]

I tried running pixi lock to update the lockfile, but it doesn't seem to work on my Mac.