This is an interesting study with an enormous sample size with results suggesting cannabis users may show mitigated cognitive decline. These findings are particularly exciting given they corroborate a small rodent literature indicating that cannabis may protect against age-related declines in cognition.

Introduction
The following statement should be immediately followed by citations of reviews and meta-analyses supporting that aspect of the statement rather than including these citations at the end of the sentence containing two different statements “While there is substantial evidence of the short-term negative effects of cannabis use on cognition and cognitive development”. I also suggest you add this meta-review of meta-analyses to the list of citations.
Dellazizzo, L., Potvin, S., Giguère, S., & Dumais, A. (2022). Evidence on the acute and residual neurocognitive effects of cannabis use in adolescents and adults: A systematic meta-review of meta-analyses. Addiction, 117, 1857-1870. doi: 10.1111/add.15764

Please include a separate list of citations to support the following statement immediately following it “…its link to age-related cognitive decline has been sparsely studied and has shown mixed results.” Again, I suggest you add this study to those cited to support this statement.
Watson, C.W., Sundermann, E., Helm, J., Paolillo, E. W., Hong. S., Ellis. R. J., Letendre. S., Marcotte, T. D., Heaton, R. K, Morgan. E. E., & Grant, I. A. (2023). Longitudinal study of cannabis use and risk for cognitive and functional decline among older adults with HIV. AIDS and Behavior. doi: 10.1007/s10461-023-04056-6

In general, the review of the relevant literature on cannabis use and age-related cognitive decline is sparse and insufficient in detail. Please include some details on study methodologies and directions of findings rather than simply referring to findings as “mixed.” If space is an issue, then discussion of rates of cannabis use could be dramatically reduced.

Methods
Please restate this range as lowest to highest mean retest interval (range = 28 years).

Please also provide information about the range of ages of the men rather than just the years of examinations “The draft board examinations were conducted in 1967-89, while the follow-up examinations were conducted in 2015-17 and 2019-22.”

Why was age of initiation recategorized into three levels? It would be more appropriate to leave this as a continuous variable.

These response options should be reorganized to appear in the order of least to most frequent “Never/Almost never, Less than once a week, Approximately once a month, A couple of times a month, Approximately once a week, A couple of times a week, Every day/Almost every day)”

The age periods used to determine persistent use are not equal and it is not clear then how number of years of persistent use was determined. In general, the following statement is unclear because the term persistent typically refers to duration. “Frequency of cannabis use was operationalized as the number of years of persistent cannabis use with persistent cannabis use being defined as using cannabis a couple of times a week or more. It may be clearer to state that “Frequent cannabis use was operationalized as using cannabis a couple of times a week or more.” The term years of persistent use could be changed to years of frequent use or regular use (i.e., years of regular use) to reflect this change.

Why was persistent use recategorized into two levels. Again, it would be more appropriate to leave this as a continuous variable.

This statement is also unclear “years of education were calculated using the nominal lengths of study times.”

Please clearly indicate which variables are predictors, which are outcomes, and which are covariates in the statistical methods subsection.

Results

Effect sizes (e.g., Cohen’s d for t-tests and R2 for regressions) should be presented along with p values especially given the enormous size of the sample. Also providing R2 values for all three models will allow the reader to better understand if the two null models are simply less powered or whether the effects (proportion of variance accounted for) are of different size.

Beta weights are standardized and cannot exceed 1. The beta symbols should be changed to lowercase italicized b (b) which is the symbol for an unstandardized beta weight. Alternatively standardized beta weights could be reported with the current symbol.

Why do you indicate the data are not shown when they are in the table? In Table 2, the reference group consisted of non-users with an unadjusted mean cognitive decline of 6.8 IQ points (SD = 9.5) (data not shown).

Discussion

The citation below provides similar evidence but in a patient population. The statement should be adjusted accordingly. “Few similar studies with similar long follow-up periods exist and these studies did not find less cognitive decline among cannabis users compared to non-users.”
Watson, C.W., Sundermann, E., Helm, J., Paolillo, E. W., Hong. S., Ellis. R. J., Letendre. S., Marcotte, T. D., Heaton, R. K, Morgan. E. E., & Grant, I. A. (2023). Longitudinal study of cannabis use and risk for cognitive and functional decline among older adults with HIV. AIDS and Behavior. doi: 10.1007/s10461-023-04056-6

The results should be discussed in the context of the animal literature that supports the present findings. See for example:
Marchalant, Y., Cerbai, F., Brothers, H. M., & Wenk, G. L. (2008). Cannabinoid receptor stimulation is anti-inflammatory and improves memory in old rats. Neurobiology of Aging, 29, 1894–1901. doi: 10.1016/j.neurobiolaging.2007.04.028.
Bilkei-Gorzo, A., et al. (2017). A chronic low dose of Δ9-tetrahydrocannabinol (THC) restores cognitive function in old mice. Nature Medicine, 23, 782-787. doi: 10.1038/nm.4311.
Sarne, Y., Toledano, R., Rachmany, L., Sasson, E., & Doron, R. (2018). Reversal of age-related cognitive impairments in mice by an extremely low dose of tetrahydrocannabinol. Neurobiology of Aging, 61, 177-186. doi: 10.1016/j.neurobiolaging.2017.09.025.

Personality differences (openness to experience) are a potentially confounding variable that may be driving the results. This could be discussed.

It is problematic that the majority haven’t continued to use cannabis. Additional information on frequency of use during different age brackets would be useful (perhaps in a table or supplemental table).