Lmaxes specified does not match number of tissues

Good question. Personally, I don’t see any reason why you can’t do a fixel-based analysis on regular b=1000 data. What it does mean though is that the simple interpretation of your fibre density values as being proportional to the intra-axonal volume won’t be as valid. The extra-axonal signal won’t be attenuated to negligible levels, so it will also contribute to the overall signal, and hence to the apparent fibre density. Note that some people would argue that we need b>6000 to fully attenuate the extra-axonal compartment, in which case this confound will also be relevant for b=3000, although to a much lower extent.

On the upside, that signal only corresponds to ~20℅ by volume in healthy white matter, and it will still be more strongly attenuated than the intra-axonal signal, so its contribution to the overall DW signal should be relatively minor.

So in my opinion, the downside of lower b-values (other than a minor loss of angular resolution) is that any differences observed in a fixel-based analysis are slightly harder to interpret: the difference may be related to changes in the extra-axonal space, rather than to changes in the intra-axonal volume… This is problematic because we’d expect intra-axonal volume to correlate very strongly with fibre density, so introducing a confound with the extra-axonal signal dilutes the ‘purity’ of that interpretation. But all that means is that you have to be a bit careful not to over-interpret changes you might observe in your analysis…