As I biohacker I do quite a bit of self-experimentation. There is an interesting ethical question about self-experimentation. Obviously people have the right to make their own decisions. Self Experimentation has a long history in Medicine and at least five Nobel Prize winners have won a prize following self experimentation. However, it is potentially dangerous and people have died.
The ethical question is whether researchers should be penalised by being unwilling to self-experiment. To that extent some US ethics committee argue against the academic publishing system accepting the results of self-experimentation. I personally, unsurprisingly, think that is wrong. There should not be a condition of employment that people self-experiment, but it cannot be right to exclude the results.
We also need to recognise that there are serious problems with animal experiments. Everything Wrong with Mouse Studies (Kinda) subtitled: Odors, magnetic fields, and even a mouse's siblings can all impact science. is an interesting article with references.
Gemini summarises it in 100 words as:
chatGPT summarises it as:
Asimov Press argues that mouse experiments often fail to replicate—or translate to humans—because lab results are highly sensitive to “hidden” variables. It opens with a cautionary example: engineered T-cells killed many mice yet didn’t trigger the predicted catastrophe in cancer patients. The essay catalogs confounders that can flip outcomes: cage location, nighttime light, bedding depth, enrichment toys, shelf stress, humidity, diet differences, microbiome variation, and even experimenter sex/odor affecting pain. It notes the scale of mouse use and low drug-translation rates, then suggests pragmatic fixes (standardize conditions, include females, raise temperature) and complementary approaches like organoids. Still, mice remain indispensable.
I think it is actually worth reading the original article. An additional question for lifespan experiments is that mice are often euthanised if they look as if they are really ill. Hence this can skew how long they live depending upon the subjective views of the animal welfare people caring for them. This is normally not reported.
C. Elegans experiments also have substantial problems. C. Elegans somatic adult cells don't divide so you cannot see the results of senescence and they have a contest with their live food (E. coli). Hence if they are fed on live food they will live longer if fed an antibiotic. C. elegans research also faces severe reproducibility issues due to the worms' hypersensitivity to environmental variables and genetic drift between labs. Additionally, their biological simplicity-lacking hearts, lungs, or adaptive immunity and thick cuticles that block drug uptake limit their ability to accurately model complex human physiology and disease.
Is an interesting video which argues strongly for the merits of N=1 experiments.
I have published a summary of this video on Rapamycin news
I am myself not sure how blinding and randomisation can work in the context of this and I think there are arguments for keeping detailed records of n=1 multiple variable experiments with a view to finding interventions that move the needle in the right direction. The difficulty of combining interventions in large trials makes this an important part of discovery.
The ethical question is whether researchers should be penalised by being unwilling to self-experiment. To that extent some US ethics committee argue against the academic publishing system accepting the results of self-experimentation. I personally, unsurprisingly, think that is wrong. There should not be a condition of employment that people self-experiment, but it cannot be right to exclude the results.
We also need to recognise that there are serious problems with animal experiments. Everything Wrong with Mouse Studies (Kinda) subtitled: Odors, magnetic fields, and even a mouse's siblings can all impact science. is an interesting article with references.
Gemini summarises it in 100 words as:
The article by Niko McCarty on Asimov Press examines the "reproducibility crisis" in biomedical research. It argues that despite rigorous attempts to standardize laboratory conditions, "hidden variables" make consistent results nearly impossible.
Key Points
- Environmental Sensitivity: Mice are profoundly affected by subtle factors like cage bedding, ambient temperature, lighting, background noise, and even the gender of the human handler.
- Genetic Drift: Even "identical" inbred strains accumulate mutations over time, creating genetic divergence between labs.
- The Result: These uncontrolled variables mean a treatment that works in one lab often fails in another, wasting resources and slowing medical progress.
chatGPT summarises it as:
Asimov Press argues that mouse experiments often fail to replicate—or translate to humans—because lab results are highly sensitive to “hidden” variables. It opens with a cautionary example: engineered T-cells killed many mice yet didn’t trigger the predicted catastrophe in cancer patients. The essay catalogs confounders that can flip outcomes: cage location, nighttime light, bedding depth, enrichment toys, shelf stress, humidity, diet differences, microbiome variation, and even experimenter sex/odor affecting pain. It notes the scale of mouse use and low drug-translation rates, then suggests pragmatic fixes (standardize conditions, include females, raise temperature) and complementary approaches like organoids. Still, mice remain indispensable.
I think it is actually worth reading the original article. An additional question for lifespan experiments is that mice are often euthanised if they look as if they are really ill. Hence this can skew how long they live depending upon the subjective views of the animal welfare people caring for them. This is normally not reported.
C. Elegans experiments also have substantial problems. C. Elegans somatic adult cells don't divide so you cannot see the results of senescence and they have a contest with their live food (E. coli). Hence if they are fed on live food they will live longer if fed an antibiotic. C. elegans research also faces severe reproducibility issues due to the worms' hypersensitivity to environmental variables and genetic drift between labs. Additionally, their biological simplicity-lacking hearts, lungs, or adaptive immunity and thick cuticles that block drug uptake limit their ability to accurately model complex human physiology and disease.
Is an interesting video which argues strongly for the merits of N=1 experiments.
I have published a summary of this video on Rapamycin news
chatGPT summary of transcript: Arguments Justifying N=1 Experimentation
Based on the interview with Dr. Gordon Guyatt, the "Father of Evidence-Based Medicine," N=1 trials (single-patient randomized crossover trials) are conceptually placed at the top of the evidence hierarchy for individual decision-making. The key arguments include:
- Individual Specificity vs. Group Averages: Standard Randomized Controlled Trials (RCTs) only reveal the average effect of a treatment on a population. Due to "heterogeneity of treatment effect" (biological differences between people), an average result cannot guarantee how a specific individual will respond. N=1 trials determine definitively if a treatment works for that specific patient.
- Methodological Rigor: Unlike casual "try it and see" self-experimentation, N=1 trials employ randomization and blinding. This rigor eliminates common errors caused by the placebo effect, confirmation bias, and "regression to the mean" (natural fluctuations in symptoms), ensuring that observed improvements are actually due to the treatment rather than coincidence or expectation.
I am myself not sure how blinding and randomisation can work in the context of this and I think there are arguments for keeping detailed records of n=1 multiple variable experiments with a view to finding interventions that move the needle in the right direction. The difficulty of combining interventions in large trials makes this an important part of discovery.
Comments