Tools to Enhance Working Memory & Attention

The episode opens by defining working memory as a special category of memory: the ability to hold small amounts of information in mind for short periods of time. The host, who introduces himself as Andrew Huberman, a professor of neurobiology and ophthalmology at Stanford School of Medicine, frames the topic around its close relationship to attention. The claim is that anyone interested in better focus and attention can benefit from understanding what working memory is and what can be done to improve it.

The stated plan has three parts: explain what working memory is, including some of its underlying biology; describe tools to improve it; and compare it with long-term and short-term memory. The host expresses confidence that this comparison will help listeners both develop better focus and commit certain kinds of information to short- and long-term memory. He also notes that the podcast is separate from his teaching and research roles at Stanford.

To set up the contrast, the discussion starts with long-term memory, which is what most people mean by the word memory. Examples include remembering the capitals of states or countries, the continents, directions from one place to another, and even one's own name.

Long-term memory is divided into two components. Declarative long-term memories are things we can declare, such as facts about ourselves, others or the world. Procedural long-term memories are the action steps behind procedures like riding a bicycle or driving a car. Once learned, these may run without conscious awareness, but even after they become reflexive they are still stored in long-term memory.

On the biology, the hippocampus is singled out as a key structure within a larger network that is described as absolutely essential for forming and storing long-term memories. The name means seahorse in Latin, and there is one on each side of the brain, hence the plural hippocampi. People with damage to the hippocampus have trouble forming or accessing long-term memories, and sometimes both. The host points to an earlier full episode on long-term memory and says a few of its tools will come up later.

Working memory also has to be separated from short-term memory. Short-term memory is described as a temporary bank for information that may or may not be passed into long-term memory. Learning English or learning to write, for example, required the relevant information to be held in short-term memory first. These memories are said to last somewhere between a few minutes and potentially a few hours, maybe a little longer. Only a certain percentage of them makes it into long-term memory.

The argument here is about progressive filtering. Whether someone hears a podcast or attends a lecture on cognitive material or a physical skill, they will remember more immediately afterward than they will a week later. The host bases this on what he calls "probably millions of scientific papers and studies." Only a small percentage of what we perceive enters short-term memory, and only a fraction of that reaches long-term memory.

The transfer from short-term to long-term memory is said to occur partly within the hippocampus. The host adds a less familiar point: some memories we think of as long-term are actually distributed into the neocortex, the outer portion of the brain. He says the names matter less than the broader theme, which recurs throughout the podcast. Rarely, if ever, does something happen in one brain location. Memory formation is a network phenomenon in which many areas collaborate, pass information along, and store it in a distributed way.

A key difference between working memory and short- or long-term memory is that the latter almost always involve neuroplasticity, defined as the nervous system's ability to change in response to experience. The host criticizes popular coverage for treating neuroplasticity as one thing, when in fact there are many types.

The first type is long-term potentiation (LTP): connections between neurons get stronger when those neurons repeatedly fire very close together in time. The host corrects a common misattribution here. The phrase "fire together, wire together" is often credited to Donald Hebb, a psychologist in Canada who did write about neuroplasticity, but the host says it was actually coined by Carla Shatz, his colleague at Stanford. In practical terms, when a small group of neurons (anywhere from two to 2,000) is active close together in time and the neurons have physical access to one another, the result is often, though not always, LTP. Afterward they communicate more easily through electrical and chemical activity. The host compares this to removing a wall between two people in conversation.

The second type is long-term depression (LTD), the inverse of LTP. It is the weakening or removal of synapses between neurons. The host stresses that the name has nothing to do with depression as a psychiatric or psychological symptom. Both LTP and LTD are described as critical for many kinds of learning and for forming short- and long-term memories. They are also involved in removing memories, meaning literal forgetting. The host adds that these are not the only mechanisms: there are dozens or more, including spike timing-dependent plasticity and paired pulse facilitation.

The third type, neurogenesis (the formation of new neurons), gets a more qualified treatment. It is robust in the developing nervous system of animals and humans. Adult neurogenesis is real and exciting and has drawn heavy press coverage, and it has been reported even in people well into their 80s and 90s. Still, the host argues that the total amount in the adult human brain is "infinitesimally small" as a mechanism for plasticity and learning compared with LTP and LTD. He notes some evidence of new neurons in the dentate gyrus of the hippocampus. He also acknowledges a decades-long controversy over how much neurogenesis occurs and whether it continues after puberty. His conclusion: when someone learns a fact or a motor skill, most of the storage comes from strengthening some connections and weakening others, not from new neurons.