Very insightful and informative manuscript. I have a couple of comments that I would like the authors to address in their rebuttal letter and manuscript (please track changes).

1. Introduction 1. could you please provide some insight on what led you to hypothesise that sleep/chronotype would be a greater impact on brain/WM integrity. 2 why did you include chronotype and subjective vs objective sleep quality (e.g. EEG, sleep study) 3. did you approach your diagnosis approach within an RDoC approach or who/how were psychiatric disorders diagnosed?

2. please define what you mean by morningness/eveningness. I can guess what you mean but it would be good to define these terms and provide some background on how they were previously related to white matter integrity in previous research papers e..g sleep disorders? neurodevelopmental conditions involving "glitches" in the circadian rhythm.

3.clinicians/researchers are aware that sleep/chronotype is often not harmful per se, it is certainly a red flag if diagnosis has not been finalised and can make illness management more challenging. as you are looking for a link between sleep and white matter integrity I am wondering if you are suggesting that a specific chronotype would a marker for a worse illness progression? it may be helpful to define in your hypotheses if there is simply a relationship or a direction in your cause/effect.

1. your audience certainly knows what FA, RD, AD mean but it may be helpful to provide some background on why you think that these specific markers would be informative given the topic, what do we know from previous ML research in terms of sensitivity/accuracy of these parameters. also background on how poor sleep or chonotype affect myline? cerebrovascular functioning? may he helpful so that we can link these parameters to sleep.

2. I am intrigued as to why you selected subjective sleep quality. Often in psychiatry sleep study suggest that clients sleep the right number of hours but often report they sleep less or have poor sleep (subjective/objective discrepancy). this may be associated with ruminations, worries, intrusive thoughts etc. could you please comment on this.

3. I noticed that you selected a number of ML models. could you please define why you picked these? any advantage of one over the other?

4. I also noticed the number of white matter tracts you selected. any comments as to why you picked all of them and not specific ones based on previous literature or known link between sleep, neurobiology/neurochemistry/tracts?

5. could you please include a matrix of accuracy/sensitivity for your models? I am trying to figure how likely it would be for your data to predict membership to either "a control" or an individual experiencing bipolar/MDD based on your data.

6. in your figures/tables could you please include a caption to highlight relevant findings? and what we are looking at? readers often look at data first and would benefit from having an immediate idea of what visualisations convey.

7. could you please include some sociodemographic data for your participants, including education, professions, general functioning levels (FAST? GAF? or similar), and estimates of intellectual functioning (IQ? NART? education?)

8. I will read your discussion again once I read your revisions. thank you!

The authors addressed my concerns in a satisfactory manner.