Chapter 4 Chapter Summary

Chapter Overview:

The primary goal of this chapter is to present the neuroscientific theories that explain how we form and maintain long-term memories.

Our understanding of the brain’s role in memory originated from the case of H.M. (Henry Molaison). H.M. had severe epilepsy that did not respond to drug treatment. As a result, doctors removed his hippocampus and most of the medial temporal lobe (MTL). While the procedure was successful in reducing his seizures, it left him with antegrade amnesia (the inability to form new memories) for the remainder of his life. Despite this deficit, H.M.’s intelligence and language skills were still intact, he could learn new skills and he could hold information in his STM. This now pivotal case illustrated that different parts of the brain were involved in different memory processes.

Even before the case of H.M. researchers were theorizing about neural connections in long term memory. The most influential of these was Hebb’s plasticity which proposed that neural connections could change with experience so that some connections became stronger or more responsive. This plasticity is the result of Hebb’s postulate that states that if a neuron persistently causes another to fire, metabolic changes will occur to make the postsynaptic neuron more responsive to the presynaptic neuron. Hebb’s postulate does a good job of explaining classical conditioning.

While Hebb’s explanations were sound, empirical evidence to support them did not come until the 1970s. Bliss and LØmo (1973) measured an increase in population spikes in the perforant path of the hippocampus after repeated exposure to tetani. Because the spikes appeared well after the last impulse was present, they dubbed the phenomenon long-term potentiation (LTP). By illustrating the presence of LTP, Bliss and LØmo also provided concrete evidence for Hebb’s plasticity.

The metabolic changes that explained Hebb’s postulate was illustrated years later by Rao and Finkbeiner (2007). When the neurotransmitter glutamate is initially released into a synapse in the hippocampus, it will bind to both NMDA and AMPA receptors. However, because the NMDA’s calcium receptors are blocked by magnesium ions, nothing happens. The binding of glutamate to the AMPA receptors however will release just enough sodium ions to cause a small depolarization, which dislodges the magnesium ion and opens the NMDA’s calcium channel. The influx of calcium ions creates additional AMPA receptors on the postsynaptic membrane thus allowing the neuron to respond more rapidly to glutamate.

Morris et al. (1986) demonstrated that NMDA receptors in the hippocampus were critical in spatial memory. When the NMDA receptors in rats were inhibited, they showed no memory of a desired platform location in a Morris Water Maze.

Since H.M.’s case, it has been assumed that the role hippocampus was the formation of new memories. However, some research has questioned this assumption. Most convincing of the research were those using injections of ibotenic acid to lesion the hippocampus (thus ensuring that no other areas of the brain were damaged). These studies show that memories can still be formed without the hippocampus.

Instead, research has focused on the role of the hippocampus in spatial processing, specifically in the formation of cognitive maps. The most famous of these studies compared the hippocampus of London taxi drivers (who need their spatial memory to do the job) to bus drivers and found greater grey matter in the posterior hippocampal region in taxi drivers and that this increased with years of experience (Maguire et al. 2006).

Other research has shown that the hippocampus is essential in relational blinding—the process by which different memories of the same object are linked together. Relational binding is necessary for us to recognize an object from any viewing angle.

Memory researchers have been debating two anomalies: First, why is the hippocampus and the MTL involved in the formation of new memories and the retrieval of recent memories, but not in the retention and recall of older memories? And second, how does LTM stay plastic enough for the integration of new knowledge, but at the same time maintain the stability necessary retain previous memories. The two-stage model seems to be able to explain both. In this model, new memories are transiently encoded into a temporary store (processed by the hippocampus and the MTL) and later transferred into a long-term store (which relies on the neocortex) or are forgotten. This transfer into the long-term store occurs gradually during slow-wave sleep (SWS) in a process called memory consolidation. It is necessary to consolidate memories during sleep when there is no external stimulation because it uses the same processes as incoming stimuli. This theory also helps to explain why humans need to sleep.

In the two-stage model of memory, the EEG activity measured during SWS is actually measuring the process of consolidation. Reactivation of memories during sleep redistributes them to the cortex where they become more stable.

Obviously, not all new memories become consolidated into LTM. Some research shows that more slow-wave activity occurs if you know you will need the newly processed information in the future. This suggests that students should study the night before a test and then get a good sleep. Research in this area has also demonstrated that sleep is important in remembering actions and enhances our ability to make and carry out future plans.

The multiple-trace theory is an alternative to the two-stage theory. This theory suggests that multiple episodic memory traces will consolidate into generic semantic memory traces. The episodic memories are formed by the hippocampus and stored there but the more generic, context-free memories are stored outside the hippocampus.

Both the two-stage and multiple-trace theories support the process of consolidation, where labile or changeable memories are transferred over the course of weeks or years into more stable memories. Before consolidation, memories are vulnerable to interference. However, research has recently demonstrated that consolidated memories become changeable, and potentially modified, during retrieval. The process where static memories are retrieved, become malleable and then returned to the static state is called reconsolidation. Interestingly, reconsolidation is context-dependent so that it only occurs when the context at learning and reactivation is similar.

While people are generally quite good at retrieval, neuroscientists are curious about how we can activate specific representations and not confuse them with similar others. Research by Polyn et al. (2005) used multi-voxel pattern analysis (MVPA) to tease out similarities in fMRI signal data during learning and retrieval. Their research suggests that intentional retrieval occurs when our brain activity returns to the state it was during encoding.

Learning Objectives:

Having read this chapter, you will be able to do the following:

1. Explain the impact of the case of H.M. on the understanding of the neuroscience of memory.
2. Recite Hebb’s postulate and explain how Hebb’s postulate was proven to be true.
3. Describe the experiment that first demonstrated long-term potentiation.
4. Explain how the types of receptors found on neurons in the hippocampus facilitate the learning of associations.
5. Provide evidence that NMDA receptors are necessary for LTP.
6. Explain how our understanding of the role of the hippocampus in memory has changed as a result of research with ibotenic acid.
7. Discuss evidence that the hippocampus is involved in the development of cognitive maps.
8. Describe the stability–plasticity dilemma.
9. Describe the two-stage model of memory and explain how brain activity during SWS may serve to transfer hippocampal memories to the neocortex.
10. Identify factors that may lead memories to be selected for consolidation.
11. Describe an experiment that showed that SWS is important for the consolidation of memory for actions.
12. Explain how the multiple-trace model differs from the two-stage model of memory.
13. Compare and contrast consolidation and reconsolidation.
14. Describe what multi-voxel pattern analysis (MVPA) indicates about brain activity during encoding and retrieval.