To think is to . . . what?

kw: book reviews, nonfiction, scientists, biographies, neuroscience

Let's see, the process goes something like this:

• Neural theta bursts occur about five times per second (the rate of a small mammal's sniffing, which is significant).
• Theta bursts open neural channels AMPA and NMPA, in sequence.
• Ions enter, inducing disassembly of portions of the cytoskeleton.
• The cytoskeleton reassembles via actin polymerization, making certain synapse bulbs larger (they are about a micron or two across, smaller than an E. coli cell).
  
• Molecules called integrins lock in these shape changes, that make room for more channel receptors.
• The greater numbers of receptors increases the likelihood this neural circuit will be retained.

And that is how a memory is made. The process is called LTP or Long Term Potentiation. The early steps take milliseconds, the ones in the middle take seconds to minutes, and the later steps take up to an hour and can persist for a lifetime.

Figuring it all out did take a lifetime, that of Gary Lynch, PhD, in his lab at 101 Theory Drive in Irvine, California. The lab's address provides the title for Terry McDermott's book 101 Theory Drive: A Neuroscientist's Quest for Memory. This is a biography not just of Dr. Lynch, but of the lab itself and the motley collection of graduate students, postdocs and others he assembled over the past thirty-plus years as he drove single-mindedly to determine the mechanism of memory.

By all accounts, Gary Lynch is a champion jerk. He is as famous for his feuds as for his science. Luckily for him, his science made him rich, in spite of a long list of offended collaborators and competitors. Nearly from the beginning, he has commercialized key discoveries, and it may be that the recent ones, including ampakines, may yield effective treatments for memory diseases such as Alzheimer's Syndrome and RETT.

As I read, I remember wondering, "What if Superman had a bad personality? Would he retain his popularity?" Lynch is like Superman with an attitude. He spent much of his life knowing he was right, frustrated by the lack of tools to prove his ideas (his lab partners had to invent most of those tools over the past thirty-plus years), and alternating between elation and despair. Even in his sixties, he wears his heart on his sleeve as clearly as any four-year-old (a man after my own heart!). Though a lot of people don't like him, very few try to deny he gets results, tons of results!

Early on, he realized that the key to the great memory abilities of the mammalian brain is the very randomness of the cortex. Instinct works like pre-stored programs, which never change. Learning changes all the time, as Lynch said, "…the thing's random. The key to understanding it is, ninety percent of the human brain uses the same basic circuitry." But this randomness has its drawbacks:

We've made a deal with the devil here. It isn't the secret of our existence that we have so much plasticity. It's that we have so much stability. These dendrites have to last a lifetime. Now, if anything happens to one of them…you can't get it back. (p 105)

Put another way, we know neuron death occurs at a steady rate. Those that encode a particular memory might last a century, but whenever they die, that memory is gone. Memories are made and committed at a low rate to begin with; we have no room to recall every little thing. What we do remember is remarkable, but it is less than a percent of what we could remember if we were like DVD recorders with an infinite number of disks to fill. So we keep only the "memorable" memories, plus an assortment of less critical ones that happened to, quite randomly, rise above the noise level long enough for LTP to fix them permanently.

It is all a testament to the contingent nature of evolution. However the first collection of randomly-assembled neurons arose, they were found useful for learning stuff, and the newly-"invented" cortex just grew and grew ever since. It must have started in the olfactory bulb (the "smell center"), where a small mammal's sniffing rate of five per second became fixed as the theta rhythm that triggers learning. Smell is the oldest of the senses that utilizes learning as a key element in its working. Other parts of the LTP mechanism were cobbled together from existing molecules, such as the ones that cause blood platelets to stick together during clotting.

One albatross around the neck of neuroscience has been the term Engram, popularized by Karl Lashley. While in neuroscience an engram refers to the physical "whatever" that make(s) up a memory, the term was co-opted by Scientology to refer to only painful memory traces that their "auditing" is supposed to erase. Actually, if you remove all your engrams you'll become a wholly blank slate, no memory, no language, no nothing. The science is thus still devoid of a good replacement term. LTP is a poor substitute because it refers to only part of the process, strictly speaking.

McDermott's book covers the whole of Dr. Lynch's career, but focuses on the past ten years, during which the most sensitive tools were developed that finally proved the LTP hypothesis. The cover of the book is dominated by an image from the microscopic evidence for synaptic rearrangement on the dendrites during learning. You'd think Lynch would be ready to rest on his laurels at that point (late 2008), but you can't get an old plow horse out of the traces quite that easily. Science never stops, and neither will he.