Longitudinal Observations on DSIP Administration: A Six-Week Personal Research Log with Commentary on Existing Literature
Posted: Sun Mar 22, 2026 12:45 pm
Following several months of reviewing the available literature on Delta Sleep-Inducing Peptide, I decided to document my own n=1 observations over a six-week period. I want to be explicit upfront: this is self-reported data from a single subject and carries all the methodological limitations that implies. I am sharing it here in the spirit of contributing to the community's collective knowledge base, not as clinical guidance.
Background on my baseline: I have dealt with what I would characterize as sleep architecture dysfunction for approximately four years, specifically difficulty maintaining slow-wave sleep as confirmed by home polysomnography equipment. My sleep tracking with an Oura ring consistently showed fragmented deep sleep cycles, averaging roughly 45-55 minutes of deep sleep per night when the target for my age cohort should be closer to 90-120 minutes according to the broader literature.
I sourced DSIP from a supplier I had vetted previously for other peptide work and confirmed purity via requested COA. This is a non-negotiable step in my research process and I would strongly caution anyone reading this to treat COA verification as foundational rather than optional.
Protocol I used:
The published literature suggests variable dosing approaches. Kovalzon and Strekalova (2006) documented effects at relatively low doses in rodent models, and extrapolating to human equivalents is genuinely difficult with this particular peptide given the crossing of the blood-brain barrier remains a debated mechanism. I settled on 100mcg subcutaneous approximately 30 minutes prior to intended sleep time. I ran this five nights per week for six weeks, taking two consecutive nights off each week to observe baseline comparison data.
Weeks one and two observations:
Honestly, this is where I made what I consider my primary methodological error. I had read that DSIP may require cumulative administration before effects stabilize, but I became impatient during week one when subjective sleep quality showed minimal improvement. I then made the mistake of stacking it with a low dose of magnesium glycinate and L-theanine, which I had not been using previously. This confounded my data considerably. I should have held all other variables constant for at minimum two full weeks before introducing anything concurrent. I am flagging this transparently because I see others in this community making similar compounding errors and it genuinely undermines the interpretive value of self-experimentation.
By the end of week two, Oura data showed modest improvement in deep sleep duration, averaging approximately 68 minutes. Whether this was DSIP, the magnesium, the theanine, or simple regression to the mean, I cannot say with confidence.
Weeks three through five observations:
This is where things became more interesting. Discontinuing the magnesium and theanine at the start of week three and maintaining only the DSIP protocol, I observed a sustained improvement in deep sleep metrics through weeks three and four. Deep sleep averaged 78 minutes across this window. More subjectively notable was what I can only describe as an improvement in sleep onset quality, meaning the transition from wakefulness felt less effortful. I was not measuring sleep latency with any formal instrument beyond subjective rating on a 1-10 scale, which is a clear limitation.
Week five introduced what I consider the most significant observation: on the two off nights, sleep quality dropped noticeably back toward baseline metrics, which suggests to me that whatever effect was occurring was not simply a learned behavior or placebo response maintaining itself. The on/off pattern tracking with quality changes is at least suggestive, though I hold this interpretation loosely.
Week six and discontinuation:
I ran week six fully, then stopped entirely. The first week post-discontinuation showed a partial return toward baseline, which had largely stabilized by week two off. I did not experience any rebound insomnia of clinical significance, which was a concern I had going into this given the limited human safety literature available. The Nakagaki et al. studies from the 1980s referenced anxiolytic properties in addition to sleep effects, and I did notice what I would describe as a marginal reduction in anticipatory anxiety during the active weeks, though again this is highly subjective and confounded.
What I remain uncertain about:
The mechanism question is genuinely unresolved for me. There is debate in the literature about whether DSIP functions centrally or peripherally or through some modulation of the HPA axis. Graf et al. conducted some work suggesting stress-regulatory effects separate from direct sleep architecture effects. If the primary mechanism is stress axis modulation rather than direct sleep induction, that changes the framing considerably and might explain why subjective results varied day to day based on baseline stress levels.
I also want to raise a question for the community: has anyone who has run DSIP protocols had access to actual lab-confirmed slow-wave sleep data rather than consumer wearable data? The limitation of Oura and similar devices for measuring sleep stages is well documented. I would be particularly interested in hearing from anyone who has done concurrent actigraphy or even formal polysomnography before and after a DSIP cycle.
One final note on the peptide's stability: I stored reconstituted solution at 4 degrees Celsius and used within the recommended window, but I found the peptide to be noticeably more degradation-sensitive than other compounds I have worked with. Whether any observed attenuation in week six reflected genuine tolerance, incomplete resensitization from off days, or degraded compound is another variable I cannot cleanly separate.
Overall assessment: cautiously interesting. Not transformative in my case, but the on/off signal was compelling enough that I intend to run a cleaner protocol in the future with better controls and ideally without the week one confounding error.
Background on my baseline: I have dealt with what I would characterize as sleep architecture dysfunction for approximately four years, specifically difficulty maintaining slow-wave sleep as confirmed by home polysomnography equipment. My sleep tracking with an Oura ring consistently showed fragmented deep sleep cycles, averaging roughly 45-55 minutes of deep sleep per night when the target for my age cohort should be closer to 90-120 minutes according to the broader literature.
I sourced DSIP from a supplier I had vetted previously for other peptide work and confirmed purity via requested COA. This is a non-negotiable step in my research process and I would strongly caution anyone reading this to treat COA verification as foundational rather than optional.
Protocol I used:
The published literature suggests variable dosing approaches. Kovalzon and Strekalova (2006) documented effects at relatively low doses in rodent models, and extrapolating to human equivalents is genuinely difficult with this particular peptide given the crossing of the blood-brain barrier remains a debated mechanism. I settled on 100mcg subcutaneous approximately 30 minutes prior to intended sleep time. I ran this five nights per week for six weeks, taking two consecutive nights off each week to observe baseline comparison data.
Weeks one and two observations:
Honestly, this is where I made what I consider my primary methodological error. I had read that DSIP may require cumulative administration before effects stabilize, but I became impatient during week one when subjective sleep quality showed minimal improvement. I then made the mistake of stacking it with a low dose of magnesium glycinate and L-theanine, which I had not been using previously. This confounded my data considerably. I should have held all other variables constant for at minimum two full weeks before introducing anything concurrent. I am flagging this transparently because I see others in this community making similar compounding errors and it genuinely undermines the interpretive value of self-experimentation.
By the end of week two, Oura data showed modest improvement in deep sleep duration, averaging approximately 68 minutes. Whether this was DSIP, the magnesium, the theanine, or simple regression to the mean, I cannot say with confidence.
Weeks three through five observations:
This is where things became more interesting. Discontinuing the magnesium and theanine at the start of week three and maintaining only the DSIP protocol, I observed a sustained improvement in deep sleep metrics through weeks three and four. Deep sleep averaged 78 minutes across this window. More subjectively notable was what I can only describe as an improvement in sleep onset quality, meaning the transition from wakefulness felt less effortful. I was not measuring sleep latency with any formal instrument beyond subjective rating on a 1-10 scale, which is a clear limitation.
Week five introduced what I consider the most significant observation: on the two off nights, sleep quality dropped noticeably back toward baseline metrics, which suggests to me that whatever effect was occurring was not simply a learned behavior or placebo response maintaining itself. The on/off pattern tracking with quality changes is at least suggestive, though I hold this interpretation loosely.
Week six and discontinuation:
I ran week six fully, then stopped entirely. The first week post-discontinuation showed a partial return toward baseline, which had largely stabilized by week two off. I did not experience any rebound insomnia of clinical significance, which was a concern I had going into this given the limited human safety literature available. The Nakagaki et al. studies from the 1980s referenced anxiolytic properties in addition to sleep effects, and I did notice what I would describe as a marginal reduction in anticipatory anxiety during the active weeks, though again this is highly subjective and confounded.
What I remain uncertain about:
The mechanism question is genuinely unresolved for me. There is debate in the literature about whether DSIP functions centrally or peripherally or through some modulation of the HPA axis. Graf et al. conducted some work suggesting stress-regulatory effects separate from direct sleep architecture effects. If the primary mechanism is stress axis modulation rather than direct sleep induction, that changes the framing considerably and might explain why subjective results varied day to day based on baseline stress levels.
I also want to raise a question for the community: has anyone who has run DSIP protocols had access to actual lab-confirmed slow-wave sleep data rather than consumer wearable data? The limitation of Oura and similar devices for measuring sleep stages is well documented. I would be particularly interested in hearing from anyone who has done concurrent actigraphy or even formal polysomnography before and after a DSIP cycle.
One final note on the peptide's stability: I stored reconstituted solution at 4 degrees Celsius and used within the recommended window, but I found the peptide to be noticeably more degradation-sensitive than other compounds I have worked with. Whether any observed attenuation in week six reflected genuine tolerance, incomplete resensitization from off days, or degraded compound is another variable I cannot cleanly separate.
Overall assessment: cautiously interesting. Not transformative in my case, but the on/off signal was compelling enough that I intend to run a cleaner protocol in the future with better controls and ideally without the week one confounding error.
