Transform Your Metabolic Health & Longevity by Knowing Your Unique Biology | Dr. Michael Snyder

Welcome to the Huberman Lab Podcast, where we discuss science and science-based tools for everyday life. I am Andrew Huberman, professor of neurobiology and ophthalmology at Stanford School of Medicine. My guest today is Dr. Michael Snyder, professor of genetics at Stanford. His lab shows why people respond differently to foods, supplements, behaviors, and drugs. Think potato spikers versus grape spikers for insulin, fibers that lower inflammation for some but raise it for others, and how GLP one drugs and even psychological interventions can impact healthspan and lifespan. The theme today is individual variability. Let’s start simple: how does what we eat impact glucose, and what does a healthy glucose response look like?

Prolonged high spikes are bad; brief spikes can be normal. Grapes can give a quick, transient rise, and strength training can spike glucose because you mobilize glycogen—mine jumps when I lift. Time in range is a useful rule of thumb: for most healthy people, stay between seventy and one hundred forty; for diabetics, seventy to one hundred eighty. Continuous glucose monitors read every five minutes, so you can see your personal responses—some spike to bananas, others to potatoes or different breads. We defined glucotypes—good control, moderate spikers, severe spikers—and time in range correlates strongly with hemoglobin A one c. Glycemic index charts do not capture this personal variability.

So GI tables are not enough—measurement is key. What about how spikes feel? For instance, oatmeal now makes me sleepy post-workout, but white rice does not. Is that a glucose thing?

Sleepiness is common with large spikes. You can feel a brief boost, then crash—pizza does that to me. Alcohol can add to it. A brisk fifteen to twenty minute walk after eating blunts spikes—most people spike to white rice, and walking lowers that. Mechanistically, active muscle pulls in glucose and burns it.

I like exercise snacks—squats, pacing, even soleus “pushups.” What about meal timing for glucose control, separate from weight loss? Is earlier better?

In our study with CGMs, smartwatches, and detailed logs, people who ate their biggest meal earlier had lower glucose; big dinners pushed it up. Starting with vegetables helped, and more sleep associated with lower glucose. Exercise timing depends on your dysregulation subtype: if you are muscle insulin resistant, more morning activity improved next-day glucose. Type two diabetes is heterogeneous—muscle, liver, fat insulin resistance, beta cell secretion defects, and incretin defects all exist, and some people are combinations. I am a type two with a beta cell release defect; I gained ten pounds of muscle and it did not improve my glucose, but a drug that promotes insulin release did. We can often infer subtypes from the shape of a CGM glucose curve after a glucose challenge. The microbiome also shapes these responses. Any exercise beats none; we have not fully separated resistance versus endurance in that timing analysis yet.

How common are these subtypes, and what about GLP one drugs—beyond glucose and appetite, are they longevity tools? And should people pair GLP ones with exercise?

We do not have precise subtype prevalence yet. Thin people can be diabetic, and some obese individuals maintain good control—many organs and pathways regulate glucose. Incretin receptors are widespread; GLP ones may enhance cognition, and longevity trials are underway. Personally, GLP one therapy took my hemoglobin A one c from eight point four to five point seven. A lower dose agent got me to about six point four, then Mounjaro to five point seven. I lost fat—from one hundred forty four to one hundred twenty eight—and feel colder. I lift daily to preserve muscle. I am a metformin non-responder. Hypoglycemia and post-spike troughs are common; CGMs reveal those and they drive fatigue.

Some people microdose compounded GLP ones; they boost GLP one hundreds to about one thousand fold—very supraphysiologic. I hear about reduced alcohol craving and cognitive benefits. If people use GLP ones, is pairing with training mainly to preserve muscle?

Yes—strength training clearly reduces GLP one associated muscle loss. Bigger studies would help, but it has worked for me.

On sleep, I try to finish eating a few hours before bed. How do evening meals affect sleep and next-day glucose? Also, the gut talks to the brain via the vagus; most do well as omnivores, but I have seen autoimmune issues resolve on strict elimination, even carnivore, which hints at different microbiomes. What about fiber types and who benefits?

Avoid eating within three hours of sleep, and walk after dinner—high evening glucose correlates with poorer sleep. Morning cortisol links to glucose, but nightly patterns vary by subtype; consistent routines and bedtimes associate with lower glucose. For glucose control, the rough contributions are twenty to thirty percent microbiome, approximately twenty percent genetics, and the rest lifestyle. Fiber is not one thing—beyond soluble and insoluble, chemistries vary widely. We tested arabinoxylan and inulin in a randomized crossover with escalating doses. Arabinoxylan generally lowered LDL cholesterol about twenty five percent; inulin helped some individuals instead. Neither changed glucose. Most people under-consume fiber—often twelve to fifteen grams per day versus twenty five for women and thirty five for men. My goal is to match fibers to a person’s microbiome and blood markers; probiotics can help but often do not colonize without a supportive community.

Food as medicine resonates. I am tempted to run a simple experiment: keep diet and training constant, add one pure fiber like psyllium for a few weeks, track LDL, ApoB, and CGM, wash out, then test inulin the same way.

That is a good approach. I would also profile your microbiome to see which enzymes can digest those fibers and, if needed, pair a fiber with the right probiotic community. At scale, with enough people logging food and CGM, we can resolve optimal combinations.

You also argue organs age at different rates. How are you measuring that, and what can people monitor now? I did a whole-body MRI as a baseline and found it useful, though it costs about two thousand dollars.

We run deep longitudinal profiling—genomes, multi-omics on blood and urine, microbiome, and wearables every few months. In the first few years, forty nine people had presymptomatic major findings. We spun out a medical version that includes a thirty five to forty minute whole-body MRI. Baselines matter because everyone has nodules—what matters is growth. I have nine nodules stable across twenty MRIs. Multivariate shifts and trajectories flag risk earlier than single snapshots. Physicians are time constrained, but a trend-based model catches issues, even early ovarian or pancreatic cancer.

Beyond CGMs and sleep trackers, what do you wear and what is useful?

I wear multiple watches that track heart rate, heart rate variability, sometimes blood oxygen and skin temperature, plus galvanic skin response, which reflects hydration and stress. Some record a single-lead ECG. Sleep stage accuracy varies by device, but wearing them overnight is great for health monitoring.

I check sleep data only every few days so I do not bias my mindset—Ali Crum’s work shows sleep scores can shape how rested we feel. Warming the sleep environment in the last two hours reliably increases my REM; I now get roughly two and a half hours with about six and a half to seven hours total.

I will try that. I do not know whether mattress-embedded trackers are more accurate than wrist devices.

What can we do to improve heart rate variability? I like the deliberate long exhale to empty lungs to engage the vagus and raise heart rate variability across day and night. Also, there are data from Stanford that positive next-day anticipation powerfully improves sleep quality; when you are excited about your life, you seem to get more high-quality sleep even if you sleep a bit less.

Exercise, lower stress, and better sleep help my heart rate variability. I am not a cardiologist, but my own heart rate variability jumped roughly twenty percent when my sleep improved. I do not currently meditate; my end-of-day bike rides and some gardening are my calm. I used to do a five-minute post-exercise meditation and should get back to it.

Back to organ aging and lifespan: how much is genetics versus lifestyle? And is approximately one hundred twenty years the cap?

Average lifespan is estimated to be about sixteen percent genetic, with a wide error bar. For centenarians, genetics may be higher—some say up to sixty percent—but lifestyle still matters a lot. Blue zones share low intake of ultra-processed foods, a more Mediterranean-leaning pattern with more fish and chicken than red meat, lots of vegetables and fiber, regular activity, strong social networks, and likely good sleep, though that is less measured.

My diabetes illustrates gene–environment interplay. A polygenic risk score flagged me high-risk, but I only became diabetic after a respiratory syncytial virus infection. We saw DNA methylation shifts across about one hundred metabolic gene promoters. In our cohort, seven people developed diabetes gradually, two had trigger events like me. Post-COVID, two to four percent develop diabetes—consistent with epigenetic stressors flipping a predisposed system. Similar trigger concepts may apply to chronic fatigue and some autoimmune diseases. We also see homeostatic weak points—like under-expressed ALS-relevant genes in induced motor neurons—where an external hit can tip the system. Evolutionary tradeoffs exist too, like sickle cell variants and malaria resistance.

I think of viral hits flicking a tilted domino, and of disease as unique combinations of missing puzzle pieces. Correlations exist, like HSV-1 and Alzheimer’s. There are monogenic exceptions like Huntington’s, but even there you see escapers.

Exactly. Even single-gene diseases can have escapers, which may teach us how to help others.

On ALS: what about SOD1 and prevention—behaviors, supplements, drugs to protect motor neurons?

I am not an ALS clinician; my role was genetics. Using new AI genome methods, we went from seven known genes to about six hundred ninety, explaining more heritability. There is no cure yet. Overexercise appears harmful in ALS, but I do not know protective behaviors that clearly prevent it.

I am intrigued by neuron protection in general. Nicotine—when not smoked or vaped—has interesting rodent data for protecting dopamine neurons in two-hit models, but it is addictive and raises blood pressure, so I do not recommend it broadly. For now, it is mostly behavioral don’ts—like avoiding head injury.

Air quality: you brought a device. How do you use it, and what are you learning about the outside linking to the inside?

I carry an air meter everywhere—P M two point five, P M ten—and during fires it spikes. We built a sampler that pulls air, traps microbes and chemicals on zeolite, then we analyze them offline by mass spectrometry and sequence biology, and correlate exposures with deep internal profiles—blood metabolites, cytokines, glucose, and more. Examples: I blamed pine, but my spring allergies track eucalyptus exposure. We detect chemicals like D E E T even in my office. Prior work links pesticides to Parkinson’s; we are building models and using mediation analysis to approach causality. A fun one: pyridine in paints associates with less fungal capture; my home was painted green with no pyridine, and I see more fungal exposure. Whether that is good or bad depends on the person. We are scaling to compare kitchens versus living rooms versus outdoors and link those to inflammatory and metabolic markers.

Microplastics and forever chemicals worry me. Some studies even found more microplastics in certain glass bottle setups due to the cap liner, at least in Europe. Measuring one’s air and water seems essential. And on blood sampling—Theranos aside—can you truly get thousands of biomarkers from a single drop?

Yes, for omics. The goal is easy, at-home deep profiling, not replacing exact clinical assays. Paper collection oxidized analytes, so we validated other media where proteins, lipids, and many metabolites stay stable. We can do proteomics, lipidomics, metabolomics from micro-samples. I sampled hourly for seven days while wearing a C G M and smartwatch and logging food and activity; we found thousands of correlations with precise dynamics—for me, insulin spikes about ten minutes after glucose in specific contexts. We also see alpha-synuclein varying with stress. If validated, it might become a useful assay to track risk and push back dementia—but it is still a hypothesis.

Acupuncture: Chufu Ma’s lab maps needle combinations to splenic output via the vagus, shifting inflammatory cytokines up or down. What did you observe personally?

At U C Irvine’s integrative clinic, we tried electroacupuncture for blood pressure and diabetes. My baseline was around one forty over low eighties. The next day after one session, systolic dropped roughly twenty-five points and diastolic to about seventy-two. After four of eight sessions, I have been holding around one eighteen to one twenty over about seventy-four. I am tracking to see durability. They added stress points too. I went in open-minded; I am clearly a responder.

You also studied immersive psychological programs like Byron Katie and Tony Robbins. What did you find?

Mental health lacks objective biomarkers, so we combined surveys with wearables and micro-sampling. Small pilots with Byron Katie and Tony Robbins showed survey improvements; Byron Katie participants also showed lower inflammatory markers. We then ran a larger Tony Robbins study—about six to seven hundred attendees and a separate control cohort—surveyed before, immediately after, one, three, six, and twelve months. Not randomized, but attendees improved significantly across anxiety, depression, and burnout; controls did not. About one hundred thirty to one hundred forty attendees did serial micro-sampling; those omics data are in progress.

I appreciate how you de-silo this space—breathwork, resistance training, acupuncture, immersive programs—while bringing rigorous genomics, proteomics, and statistics. This is how we customize health. I also see a near future where physicians run whole-body scans, labs, and wearable streams through A I to actually use these data.

We are trained in silos. My cardiology care pushed higher statins while my glucose climbed—textbook—but it ignored whole-body tradeoffs. We need a systems view across genetics, epigenetics, lifestyle, and socialization. A I is essential. Tools like January A I’s Mirror can ingest genomics, medical reports, and wearables to synthesize insights—like flagging low C D eight T cells and suggesting zinc to explore—things no single physician can juggle alone. I expect future clinicians will rely on such platforms so patients get the full value of continuous measurement.