Skip to main content

Verified by Psychology Today

Matthew J. Edlund M.D.


Dementia and Cancer: the Two-Thirds Rule

Is luck the biggest factor in health outcomes?

rawpixel at pexels
Source: rawpixel at pexels

Dementia and cancer have more in common than most people recognize. Two thirds of tumors appear to be the result of random mutations. And two thirds of dementia cases also appears to be basically random, with only about 35% responsive to lifestyle changes (so far.)

The implications of this “odd” convergence is wide. First, for diseases with difficult or ineffective treatments, prevention remains the cheapest and most effective strategy. Second, both dementia and most cancers represent failures of the body to learn and regenerate. Third, this failure to learn represents a perfect point for intersecting biological intelligence and artificial intelligence, both to understand these deadly diseases and to treat them. Fourth, if two of the world's major killers, dementia and cancer, results from causes usually beyond individual control, the idea that “people should directly pay for their own health care” condemns to death a sizable part of the population.

A lot of what kills us is ill luck.


The study in The Lancet found nine treatable or correctable factors prominent in causing dementia. The top four – mid life hearing loss 9%; failing to complete secondary education – 8%;- smoking – 5%; and failing to get early treatment for depression – 4%, all highlight the importance of innate biological intelligence to the progression of dementia.

First, mid life hearing loss. Most people usually don't ponder future dementia attending rock concerts, or blasting their customized car stereos. Perhaps they should reflect more often.

Many of his feel we are highly visual. When it comes to the sense organ turned to most often, people believe the eyes have it.

Yet hearing is critical to function and survival. Hearing is never turned off. We hear twenty-four hours a day, until we turn deaf.

And partial deafness represents a particular difficulty for the brain. So much of the information brain needs to survive comes via the continuous feed of the auditory system. If only certain parts are destroyed, , the brain appears to sometimes replace the information with its own fabricated “software,” what we know as tinnitus, or “ringing in the ears.” Tinnitus complicates most learning for the brain, further increasing the risk of depression – number four on the list of correctable dementia-provoking risk factors.

Next up is failure to complete secondary education. This is postulated to increase dementia risk by decreasing “cognitive reserve” – a polite way of saying that the more the brain learns, the better it gets at survival. Some of this increased dementia risk is thought to occur through “risky” behaviors that become less likely with greater educational levels. What it really means is health is learned.

Smoking may not seem an “learning” activity of the brain. Yet nicotine resets the entire autonomic nervous system, critical to how we learn anything, especially the unconscious learning provided by biological intelligence. By provoking atherosclerosis, including “hardening” of brain and heart arteries, the body's ability to learn diminishes.

Fourth of the list, depression can be conceptualized as the brain and body "shutting down" their learning and adaptive functions. Cognitively, we know this when watching the IQ scores of people with depression plummet as the syndrome takes hold. What has not been looked at much is how depression decreases biological learning in non-conscious arenas, which helps explain depression's capacity to exacerbate systemic illness.


The work done at Johns Hopkins by Vogelstein and others argues that the more often you divide stem cells, the more mutations you get. And that means more tumors. In simplistic terms, more replacement breeds more errors. Yet biological errors occur constantly. What happens with clinically important tumors is that the immune system fails to recognize them as foreign, or if it does fails to effectively block or kill them. These are problems of learning, just as making mistakes in regenerated stem cells represent a problem of learning when they are not corrected.

Learned Health

Dementia and cancer scare people. A lot. What is impressive is that 35-40% of that load appears to be preventable or partially preventable with what we know now. When you’re talking about diseases that will ultimately afflict the majority of mankind, that’s a big number.

Yet what of the "random" elements causing dementia and cancer? Here the convergence of biological intelligence and artificial intelligence may bear fruit.

Artificial intelligence has been dominated by logic based systems that adhere easily to the binary processes of computer science. Many researchers have been trying to use AI to understand biological intelligence, through models involving neural networks and similar schemes.

Sadly, these attempts fail to fully engage the way biological intelligence works. Biological intelligence operates across many interactive systems. Many of the information flows of these systems, even for actively studied areas like immunity, remain unknown. How the different biological information systems interact is also mainly unknown. Add to this the redundancy and multivalency of evolutionarily complex biologically intelligent systems, and it becomes clear how far AI has to go before it can reliably model even simple biological intelligence functions.

Yet there are ways to get there. To understand how tumors work, interactions between thousands of variables operating in real time may be required. To understand dementia, understanding how the brain learns and keep learning will be required.

What does this imply for relatively simple, linear models of dementia, like tau proteins and amyloid as the entire cause of the syndrome? Hopefully such simplistic models will eventually get scrapped. You need to understand more than one or two variables to figure out dementia and cancer.

Yet better understanding of how biological intelligence works will greatly advantage prevention and treatment. To do that, you'll need new and more effective AI models.

Until that time, prevention of two of mankind's greatest scourges may be far more efficient than cure.