Withania somnifera — ashwagandha — is one of the most studied botanicals in the stress and cortisol literature. Two decades of randomised controlled trials, an increasingly well-characterised mechanism, and a growing body of human data make it unusual among herbal supplements: the evidence here is not anecdote dressed as science. It is actual science, with real limitations worth naming. This article walks through what the best human trials found, what the mechanism is likely doing, and where the evidence runs thin.

What Ashwagandha Is and Isn't

Withania somnifera is a perennial shrub native to India, North Africa, and the Mediterranean, used in Ayurvedic medicine for thousands of years. Modern research interest centres on withanolides — bioactive steroidal lactones concentrated in the root. It is classified as an adaptogen, meaning it is theorised to modulate the stress response rather than sedate it.

This distinction matters. Ashwagandha does not bind GABA receptors in the way benzodiazepines do, nor does it inhibit serotonin reuptake like SSRIs. Its proposed mechanism is HPA axis modulation — specifically, reducing the cortisol output from the hormonal stress cascade at the level of signalling between the hypothalamus, pituitary, and adrenal glands. It works upstream, not at the symptom.

The two most-studied standardised extracts are KSM-66 (a root-only extract standardised to ≥5% withanolides by HPLC) and Sensoril (a root-and-leaf extract with a different withanolide profile, including a higher concentration of withaferin A). Both have human randomised controlled trial data. They are not interchangeable — different extraction methods, different plant parts, different withanolide ratios, different trial designs.

The KSM-66 Trial

The pivotal KSM-66 trial — Chandrasekhar K, Kapoor J, Anishetty S (2012) — is the strongest piece of human evidence in this space. Human RCT¹ The design: 64 adults with a history of chronic stress, randomised double-blind placebo-controlled, 8 weeks. Intervention: 300mg KSM-66 twice daily (600mg total).

The outcomes are specific and documented:

• Perceived Stress Scale (PSS): ashwagandha group −44.0%; placebo group −5.5%.
• Serum cortisol: ashwagandha −27.9%; placebo −7.9%.
• General Health Questionnaire (GHQ-28): significant improvement in the treatment arm.
• No serious adverse events in either group.

This is a well-designed human RCT with a clear, measurable primary endpoint and a plausible biological mechanism. It is not a large trial by pharmaceutical standards — 64 participants is modest — and independent replication with larger and more diverse populations would strengthen the evidence base further. But within the nutritional supplement literature, it is among the more rigorous studies available.

The Sensoril Data

The Sensoril dataset — Auddy B et al. (2008) — involved 62 adults across three doses (125mg, 250mg, 500mg Sensoril) and placebo over 8 weeks. Human study² Stress scores reduced 62–64% in treatment groups. Cortisol reduced 14.5–30.5% across doses. Sleep quality improved. C-reactive protein — a circulating inflammation marker — also reduced significantly.

These are large effect sizes. Larger than the KSM-66 trial, larger than most independent nutritional research reports. That warrants a note: Caveat — this study was conducted with support from Natreon Inc., the manufacturer of Sensoril. Manufacturer-affiliated trials carry a higher risk of optimistic effect size reporting; treat the specific magnitudes as directional signals rather than confirmed figures pending independent replication.

The direction of the findings — stress reduction, cortisol modulation, sleep improvement — is consistent with the KSM-66 data and with the biological plausibility of HPA axis modulation. The magnitude should be viewed with appropriate caution.

The Sleep Connection

A separate trial by Langade D et al. (2019) — 60 adults meeting criteria for insomnia, 10 weeks, 300mg KSM-66 twice daily — found significant improvements in sleep onset latency, total sleep time, and sleep efficiency vs placebo. Human RCT³

The mechanism for the sleep effects is not fully characterised. It may involve the triethylene glycol component of the root — a withanolide subset that appears to have sleep-inducing properties in animal models — rather than the cortisol pathway directly. The clinical implication is that ashwagandha's sleep benefits may be operating through a different route than its stress and cortisol effects, meaning the two outcomes are not necessarily correlated within an individual: someone might see cortisol improvement without sleep improvement, or vice versa.

Where the Evidence Is Thin

Good evidence on ashwagandha exists. But intellectual honesty requires naming the gaps, because the supplement market around it does not.

Long-term safety: all published RCTs are ≤3 months. Case reports of hepatotoxicity — liver enzyme elevation — exist in the literature. They are rare, and the majority resolved on discontinuation, but they are a documented signal. Caveat — anyone with pre-existing liver conditions, those on hepatotoxic medications, or those who are pregnant should consult a clinician before use. The safety data beyond 90 days is insufficient to characterise long-term risk.

Cognitive claims: some trials report improvements in memory and cognitive function. The evidence exists, but the trials are smaller, more heterogeneous, and use different cognitive endpoints — making it difficult to draw strong conclusions. Lower confidence than the cortisol and stress data.

Thyroid interactions: there is evidence that ashwagandha may increase T4 (thyroxine) production. Caveat — this is clinically relevant for anyone already on thyroid hormone replacement therapy; ashwagandha may potentiate thyroid hormone levels in ways that require monitoring.

Population diversity: the majority of trials have been conducted in South Asia. Whether effect sizes generalise across different genetic backgrounds, dietary contexts, and baseline cortisol profiles has not been systematically studied.

Seeing the Stress Load Ashwagandha Is Acting On

The cortisol story in ashwagandha research relies on blood draws at two time points — baseline and end of trial. That captures the average but misses the shape: when cortisol is actually spiking, how long recovery takes, whether the HPA axis is habitually over-activated across the day or only during specific triggers. The Sense Band captures this in real time. Electrodermal activity and heart rate variability together give a continuous window into autonomic stress state — the same physiological system that drives cortisol production.

The SenseMe app makes that data readable. Not a number to optimise, but a pattern: the stress peaks, the recovery windows, the days when the biological load is high enough that downstream effects on skin — two days later — are likely. Ashwagandha may reduce the amplitude of those stress signals. The Sense Band is how you would know — not from a blood draw at the end of 8 weeks, but from your own day-by-day signal across the supplementation period.

The combination is what interests me most: an adaptogen with real cortisol evidence operating on the same HPA axis the Sense Band is reading. Not a supplement taken in the dark, but one taken with a continuous measurement of whether the stress system it targets is actually responding. That is the version of evidence-based supplementation that becomes possible when you have hardware and software designed to track it.