Saturday, October 20, 2012

Memory Schemas: the Under-used Approach to Improving Education

I just attended a “memory schema” symposium at the annual meeting of the Society of Neuroscience. The “schema” idea is that memory of prior learning provides a framework or context for new learning. That is, new information is evaluated for relevance to preexisting schema, which may influence how readily new information transfers into memory.

The notion of schema stems originally from Harry Harlow’s ideas back in the 1940s. Harlow showed that when a monkey learns a new kind of problem, he solves it by slow plodding trial and error. However, if he has experience with a large number of problems of a similar type or class, the trial and error is replaced by a process in which the individual problems are eventually solved insightfully. For example, if you learn how to do task A, B, and C, when presented with a new task D, you might say to yourself, “I don’t know how to do this task D, but it is like task B, and I do know how to do that!” Thus, you have a leg up on learning how to do task D. The idea underlies how people become experts in a given field: their accumulated learning of various tasks provides them with a repertoire of what Harlow called “learning sets” that makes it easier to learn new things.

Few of the speakers or audience discussants seemed to be aware of this literature, and their ideas weren’t really all that new, except that the focus is now shifting to memory instead of insight. The basic idea of memory schemas is that associations among learning objects profoundly affect how easily and well a person can remember. Certainly, memory is promoted when learning objects are congruent, that is, have meaningful relationships. Sometimes, however, you can easily remember incongruent items because they are so different. These ideas are important to education, and in the panel discussion at the end of the symposium, speakers were asked to address this matter. But nobody did. And in school systems, few educators do either.

Master teachers have always known intuitively to structure meaningful relationships among learning objects. In principle, this is done by creating associations of word pairs, concepts, spatial locations, and assorted rules and principles. All these things make it easier for students to learn. The problem is that we educators don’t devote enough thought about practical ways to create structured relationships that will promote memory formation and recall. I don’t follow much of the educational research literature, but I suspect that very little of it focuses on the best way to organize the presentation of learning materials. For instance, has anybody conducted an experiment that tests how well students learn the central concepts in the U.S. constitution and its amendments, depending on how the concepts are presented? Or what’s the best way to structure learning objects in the teaching of cell organelles and their functions? Typically, in the latter case for example, a biology teacher considers each organelle in turn and spews out information on what it does. That may not be the most memorizable way to present the information. Maybe it would be better to begin with the biological needs of the cell, how those needs relate to each other, and then how various organelles fulfill those needs. In fact, I took that kind of approach in the on-line biology curriculum I wrote, but no experiment has compared the ease of learning this way versus the traditional approach. I do know from my own experience with trying to learn a little Spanish that the ease of memorizing verb conjugations was greatly affected by how I laid out the words in a table.

To return the schema symposium, the experiments reported made it clear that structured relationships of learning objects improve all aspects of learning: encoding, memory consolidation, and recall. The time has come to develop teaching strategies that exploit the brain’s preferred mode of operation.

One example is the development of a PowerPoint method I developed to create a one-flash card of learning objects that consists of mnemonic icons systematically placed in specified spatial locations. The images represent concepts to memorize, and the spatial locations create spatial relationships that promote memory of the learning objects. For details, see my e-book for students.

In more general terms, the primary task of teachers and students is to develop strategies to enrich the formation of memory schemas. This means finding ways to increase the number and congruence of associations among facts and concepts being taught. The research shows that major benefits can be expected. 


Harlow H F. 1949. The formation of learning sets. Psychol. Rev. 56:51-65.

Klemm, W. R. 2012. Better Grades, Less Effort. (e-book in all formats at

Van Kesteren, M. T. R., and Henson, R. N. A. 2012. The re-emergence of schemas in memory research: from encoding to reconsolidation. Society of Neuroscience Symposium. New Orleans.

Sunday, October 07, 2012

Behavioral Therapy Erases Bad Memories: Timing Matters

It has taken 50 years, but memory research has finally put it all together to provide practical guidance to reduce forgetting of what we need to remember and promote forgetting of useless or disturbing memories.

I have blogged before about animal studies showing that bad memories can be erased. Bad memories are often created like conditioned reflexes in Pavlov’s dogs. That is, the situational context in which bad things occur act as associational cues that help cement the memory. If the cues are repeatedly present but the bad event is not, the learned associated tends to go away.
But in both animals and humans, this “extinction” as it is called is not really permanent and the bad memories can recur.  
Memory researchers have recently discovered that when a memory is recalled, whether good or bad, there is a short time where it can be modified by new thought or experience, and then it is put back in storage (called “reconsolidation”). When this phenomenon was discovered, it raised the possibility that timing of extinction trials might influence effectiveness of treatments for anxiety. That is, better treatment results might occur if extinction is attempted during the vulnerable reconsolidation stage. In 2009, Joseph LeDoux and his colleagues demonstrated in rodents that timing of extinction trials did in fact influence the erasure of fear memory.  
This modifiable stage provides a way to treat even really bad memories, like post-traumatic stress disorder. When a soldier, for example, recalls a bomb killing a buddy, that terrible memory is subject to modification before it is re-stored. A typical modern treatment for PTSD is to inject an anxiety-reducing drug just before the bad memory is triggered that interferes with reconsolidation of memory. This process may have to be repeated many times before the bad memory is finally gone. Now a new study from the Uppsala University in Sweden has shown that bad memories can be erased without drug.
Investigators created bad memories in human volunteers by giving them an electric shock each time a certain picture was flashed on a computer screen. They repeated this experience 16 times to establish a conditioned fear response. The next day after such training, the subjects were brought back into the test room, and the fear-of-shock memory re-triggered by showing the picture that had been associated with shock. This was repeated eight times without associated shock, as a way to produce extinction. Half of the subjects received their extinction treatments at 10 minutes later in which the fearful stimulus was repeated without any shock. The other half of the subjects were given the same extinction treatment but delayed six hours, when it was presumably too late to interfere with reconsolidation.
To measure the amount of fear evoked by later presentations of the picture, investigators objectively measured the amount of fear, using a skin conductance test that measured essentially how sweaty the palms were. Signs of fear were absent in the group given extinction trials at 10 minutes when reconsolidation was still in progress. But signs of fear persisted in the six-hour group.
To pursue questions about what was happening in the brain, investigators used brain imaging, and particularly noticed activity changes in the amygdala, a structure deep within the brain that is hyperactive in the presence of fear memories. On the third day, all subjects were brought back to the lab and brain scans run when the fearful image was shown. In those subjects in the six-hour group, activity in the amygdala predicted whether signs of fear (skin conductance) would return. No such prediction occurred in the 10-min group. In other words, people who lost their fear memory, as indicated by skin sweating, also lost the signs of the memory in the amygdala. Similar effects were seen in the network of other brain areas linked to the amygdala in the processing of fear memories.
None of this should have been surprising. Back in the 1960s, I and many others conducted studies in animals that showed memory of a learning event depended on what happened shortly after the learning. We knew that this short window of time was vulnerable to other mental events that could prevent memory consolidation. Implications for education were obvious: multi-tasking, for example, introduces mental events that interfere with memory consolidation. But one wonders why it took science 50 years to apply what we knew about consolidation to the treatment of anxiety disorders. The key was the recent discovery that recall of a memory puts it back in the vulnerable position of having to be reconsolidated.
Agren, T. et al. (2012). Disruption of reconsolidation erases a fear memory trace in the human amygdala. Science. 337 (6101): 1550-1552.