Sunday, August 25, 2013

Young neurons in an old brain

At the age of 88, my seemingly healthy and vigorous father suddenly died.  He had commanded an army, lived through a revolution, met kings and presidents, and through it all raised a family.  And then one night, all those memories, all those experiences, vanished.  Coming home from the funeral I realized that a library, stacked with history books, had burned to the ground. 

When we experience something, it can become a memory.  But what is this memory?  What is its neural substrate?  Can someday memories that we store in our brain be read out and stored in a machine?  Is there any hope that the library can be saved from the fire that consumes us as we die?

Standard model of memory

Our model of memory today is one of synaptic plasticity.  When we experience something, the neurons that are engaged by that experience produce electrical activity, and that activity can alter the strength of synapses that connect them to other neurons.  The electrical activity can also result in growth of new synapses.  Together, this altered strength of connectivity in an existing network of neurons is thought to be the basis of memory.  So in principle, if one could measure the strength of each synapse, and model the functional properties of each neuron, then one has a representation that approximates the state of brain of an individual.  The lifetime of memories and experience are within this representation. 

The problem, unfortunately, is that this concept of memory relies on the assumption that neurons themselves are fixed nodes, whereas the connections (that is, the synapses) are the changing components through which memories are stored.  This assumption, as it turns out, is false.  New neurons are born every day, and the human brain, even in old age, adds and subtracts nodes to the network.

Finding a neuron’s birthday

Between 1955 and 1963, there were numerous above ground tests of nuclear weapons.  With every explosion, the amount of isotope 14C was elevated in the atmosphere.  In 1963, there was a treaty that banned such tests, and since then the atmospheric level of 14C has declined because of uptake by plants.  This uptake takes place as 14C in the atmosphere reacts with oxygen to make CO2, which is then taken up by plants in photosynthesis. 

When we eat plants, or eat animals that feed on plants, the 14C is transferred to our body.  Once transferred to our body, 14C becomes part of the DNA of new born cells.  This happens when a cell divides and makes a copy of its chromosomes.  The copying process integrates the 14C into the newly made genome, making it so that by looking at the concentration of 14C in a cell’s DNA, and comparing it to the atmospheric DNA, one can tell when that cell was born. 

Kristy Spalding, Jonas Frisen, and their colleagues used this idea to find the birthday of neurons in the human brain.  In their study, they examined brains of people who had died between 2000 and 2012.  These people had had their brains preserved during autopsy, and so their brain could be studied. 

They focused their efforts on the neurons in the hippocampus region of the brain, a location that is critical for formation of new memories.  The hippocampus is the place in our brain where we form autobiographical memories, i.e., the kind of memories that describe places and people that we have met, events that have taken place in our life, etc.  Spalding and colleagues asked, how old are the neurons in the hippocampus of a person who was 30 years old when she died?  You might guess, well, the neuron is probably close to 30 years old.  But that assumes that all neurons are born soon after birth.  Strikingly, Spalding and colleagues found that the neurons were much younger than the person.

Neurons are much younger than the age of the person

The authors found that for a 20 year old, the average age of neurons in the hippocampus was 18.  For a 40 year old, the average age was 29.  For a 60 year old the average age was 37.  Remarkably, for an 80 year old, the average age of hippocampal neurons was 40!  

So the average neuron in the hippocampus of an 80 year old has been around only long enough to experience the last 40 years.  It cannot ‘remember’ anything from the first half of life, because it was not around to experience it.

Therefore, there is substantial neurogenesis throughout life in the hippocampus.  In fact, the rate of neurogenesis showed only a modest decline with aging.  They estimated that each day, 0.004% of the neurons in the dentate gyrus of the hippocampus die and are replaced with new ones.

Now it is possible that neurogenesis in the hippocampus is especially high, and other parts of the cerebral cortex may not have such a high turn-over.  But the relative youth of the neurons in the hippocampus raises a fundamental question:  what is memory if neurons are eliminated and replaced on a daily basis?

Richard Feynman, the celebrated physicist, during a lecture in 1955 to the National Academy of Sciences, described the basic problem:

“The radioactive phosphorus content of the cerebrum of the rat decreases to one half in a period of two weeks.  Now what does that mean?  It means that phosphorus that is in the brain of the rat, and also in mine, and yours, is not the same phosphorous as it was two weeks ago.  It means the atoms that are in the brain are being replaced: the ones that were there before have gone away.  So what is this mind of ours: what are these atoms with consciousness?  Last week’s potatoes! They now can remember what was going on in my mind a year ago, a mind which has long ago been replaced.  To note that the thing I call my individuality is only a pattern or a dance… The atoms come into my brain, dance a dance, and then go out--- there are always new atoms, but always doing the same dance, remembering what the dance was yesterday.”

The problem in neuroscience is to understand how to read this dance.  If we could, then in principle it should be possible to record and preserve our experiences, so that when we die, the library will remain standing.

References
Spalding KL et al. (2013) Dynamics of hippocampus neurogenesis in adult humans.  Cell 153:1219-1227.
Feynman RP (1988) What do you care what other people think? Further adventures of a curious character.  Bantam Books, page 244.

3 comments:

  1. So I aak you this question: why do I remember being born?

    ReplyDelete
  2. Most of us do not remember being born, or much of anything from the first few years of life. The standard reason given for this is that the hippocampus, the region presumably critical for initially forming autobiographical memories, and the cerebral cortex, the region presumably critical for serving as the eventual storage site, are poorly developed at birth. But that doesn't mean that in infancy our brain is not forming memories. For example, there are memories that babies form when they are in the womb: a 3 day old infant can discriminate between the sounds of his mother and other women, and when given a choice, prefers to hear the sound of his mother. Perhaps we do not usually remember these early experiences because by the time we reach adulthood, there is so much neuronal turnover in the hippocampus. It is possible that if this turnover is not 100%, then some fraction of the very early experiences are still remembered.

    References
    Anthony J. DeCasper and William P. Fifer (1980) Of human bonding: Newborns prefer their mothers' voices. Science 208:1174 - 1176.

    ReplyDelete
  3. "Now it is possible that neurogenesis in the hippocampus is especially high, and other parts of the cerebral cortex may not have such a high turn-over. But the relative youth of the neurons in the hippocampus raises a fundamental question: what is memory if neurons are eliminated and replaced on a daily basis?"

    There's no paradox. Nearly all of the neurogenesis is in fact limited to the dentage gyrus, a subregion of the hippocampus. This fact is buried kind of deep in the paper you cite (more deeply buried than it should be, I think), and it's a generally accepted fact about neurogenesis now (with the exception of the newborn neurons of the olfactory system). If you subscribe to the idea that neurons in a network that support memory are fixed nodes, then you're just talking about neurons in the other two main sub-regions of the hippocampus. And in fact, those subregions (especially CA3) are the ones that modelers like to imagine are the substrate of memory. Empirically, those CA3 nodes are very fixed - as I think you would conclude they must be.

    If average neuron age being significantly lower than the human's age seems impossible to square with the idea that all neurogenesis is happening in the dentage gyrus - it's simply because there are so many more neurons in dentage gyrus than in the rest of the hippocampus. The cell type there (granule cells) are just tiny and more numerous - who knows why.. This is how it's possible that (surprisingly) half on the neurons in a human brain are in the much smaller cerebellum - cerebellum is packed with these tiny granule cells (50%, if I remember my brain trivia right).

    Some models use the turnover of cells in dentate gyrus as a kind of time-stamp for memories during their formation. .. I don't know though. I'm very fond of the camp that denies LTP as the foundation of memory, anyway. http://www.ncbi.nlm.nih.gov/pubmed/10097007

    I do like Feynman's musing about constant atom turnover though :) It gives me hope for this notion of uploading the personal library. If the atoms don't need to be the same, the neurons don't need to be the same... maybe you could simulate it all in a big computer. I couldn't agree more - what a horrible waste it is, when so many years of experience and information disappear. Sorry for your loss.

    Thanks for your post! Thought provoking.

    ReplyDelete