286‒Journal club with Andrew Huberman: light exposure on mental health & an immunotherapy for cancer

KEYPOINTS

• 01

  Outdoor light is far brighter than indoor light, and the body clock adds up photons over time

  Huberman notes that midday sun is usually over 100,000 lux and can reach 300,000. Even a bright department store is closer to 4,000–6,000 lux. A cloudy day can still be far brighter than indoors even when it doesn't look it. Because the brain's master clock sums photons over time, an overcast 8,000-lux UK winter walk still counts, as long as you aren't wearing sunglasses that filter much of the light.

• 02

  Low-angle sunlight at sunrise and sunset is the key circadian signal

  Specialized retinal cells compare short-wavelength (blue) with long-wavelength (orange/red) light. That contrast is richest when the sun is low. Huberman suggests about 10 minutes of sunlight in the eyes early in the day and at least 10 in the late afternoon, more on overcast days, and as much safe daytime light as you can get. Looking toward a low sun carries little cataract risk (blink as needed). Sunglasses are fine at midday, but he advises against wearing them for morning light.

• 03

  The study: about 86,000 UK Biobank adults wore wrist light sensors

  Participants, mostly in their 50s and 60s, wore wrist sensors measuring light from 470 to 650 nm. Researchers split them into quartiles of daytime and nighttime light exposure and compared these with psychiatric diagnoses. Models adjusted step by step for age, sex, ethnicity, photoperiod, employment, physical activity, sleep and more. The two hypotheses: more daytime light means lower psychiatric risk, and more nighttime light means higher risk. Both held up.

• 04

  More nighttime light tracked with depression and self-harm

  Compared with the lowest quartile, depression odds rose about 20% in the third quartile of nighttime light and about 25% in the fourth. Self-harm was flat across the first three quartiles, then jumped about 30% in the top 25%. PTSD (with nighttime light) and psychosis (inversely, with daytime light) also stood out. Generalized anxiety and bipolar disorder were weaker and noisier.

• 05

  How to read an odds ratio

  An odds ratio compares the odds of an outcome in two groups: subtract 1 for the percent change, so 1.78 means 78% higher and 2.0 means 100% higher. If the 95% confidence interval doesn't cross 1.0, the result is significant without needing the P value. Peter also checks whether a study is sensibly powered, and stresses clinical relevance, not just statistical significance.

• 06

  Reverse causality is the big caveat

  Depressed people may stay up late on screens and avoid going outside, so behaviour could drive light exposure rather than the other way round. Peter's example is diet soda: it correlates with obesity probably because people gaining weight switch to it. Huberman, while admitting bias, guesses 65–75% of the effect comes directly from light. Peter, using the Bradford Hill criteria (including the dose-response across quartiles and strong mechanisms), would say about 80/20.

• 07

  Cancer evades the immune system even though most solid tumours are visible to it

  Lung, breast, prostate, colorectal and pancreatic cancers cause more than half of US cancer deaths. Peter says at least 80% of solid tumours present antigens that the immune system can recognize, yet they survive by hijacking the brakes on T cells. Checkpoint inhibitors remove brakes such as CTLA-4 and PD-1, rather than pressing the accelerator. The scientists who discovered these brakes won the Nobel Prize, which Peter recalls as 2018.

• 08

  The ipilimumab phase III trial in metastatic melanoma

  Fewer than 700 patients with advanced melanoma who had failed standard therapy and had perhaps 6–12 months to live were randomized 3:1:1. The groups were anti-CTLA-4 plus a gp100 peptide vaccine, anti-CTLA-4 alone, or gp100 alone as the control. More than 70% had metastases in their organs and more than 10% in the brain. Median survival was about 10.0 and 10.1 months on the drug versus 6.4 months on gp100 alone. Peter focuses on the right-hand tail of the survival curve, where a minority of treated patients survived for years.

• 09

  Taking the brakes off brings autoimmune side effects

  Almost everyone reported some adverse event (about 97–98% in all arms), but immune-related events were about 60% on the drug versus about 30% on the control. Grade 3–4 immune events were 10–15% versus 3%. The gut is hit most often, with colitis serious enough to need IV fluids, and skin effects such as vitiligo also occur. These events are often reversed with corticosteroids, sometimes after stopping the drug.

• 10

  Why melanoma responds, and why Peter calls immunotherapy the best hope

  Most cancers carry roughly 40–50 mutations. Melanoma and kidney cancer carry hundreds, which gives T cells more targets ('shots on goal'). Peter argues that overall survival for metastatic solid tumours has barely moved since the 1970s apart from immunotherapy. He imagines a future of banking young T cells around age 30 to expand later, since T cells taken from inside tumours tend to be too exhausted to multiply a thousandfold.