Memory formation is often prevented when one event follows too soon after an initial learning event. It is also true that memory of initial learning events can be blocked if you try to learn two things at once. In fact, learning may be disrupted for both things.
In a recent test of this phenomenon, a group of 29 people (17 to 30 years of age) was trained to discriminate two sound pips that differed in length by a fraction of a second. In one group of subjects, the training occurred consecutively, which ordinarily produces some inefficiency with learning because the second task interferes with remembering the first. In this study, some learning did occur, in spite of the sequential tasks. However, results from another group of subjects revealed that when practice on the two tasks was interleaved, there was no learning on either condition.
This indicates that acquistion (initial learning) is vulnerable to multi-tasking, perhaps even more so than when learning of one task is followed too soon by another learning task. In other words, multi-tasking can interfere with initial learning, just as it does with formation of memory.
Source: Banai, K. et al. 2010. Learning two things at once: differential constraints on the acquisition and consolidation of perceptual learning. Neuroscience. 165: 436-444.
Showing posts with label learning. Show all posts
Showing posts with label learning. Show all posts
Sunday, March 7, 2010
Friday, January 15, 2010
Learning Versus Memory
Versus? Learning and memory are different, but like two sides of the same coin. What is the difference? Learning is the acquiring of new information or skills. Memory is the remembering of what was learned. You can’t have memory without learning. You can, of course, have learning that you forget.
Learning involves at least four major processes. It all begins with registering new information. This is the stage when information is detected and encoded in brain. Paying attention obviously facilitates the registration process. Multi-tasking can create an information overload in which much of the information never gets registered. Example: a car driver who is all wrapped up in a cell phone conversation may not realize she just ran a stop sign or cut off the driver behind her in the next lane. Another example comes with reading. Reading comprehension (learning) depends heavily on the eyes actually seeing each cluster of words. The reader needs to focus on words, not letters, and needs to think about what the words mean. Likewise in images, what you learn from an image depends on the details in it that you actually notice and think about.
Next is integration. The brain likes to classify, categorize, and organize its information. Thus, new information has to be fitted into existing learned schema. This is the stage where associations are made with existing memory. Brains are really good at detecting and constructing relationships. If a given relationship is not immediately obvious, the brain may figure it out and remember it. Constructing such relationships is an integral part of the learning process.
Associations can be constructed subconsciously. If two things happen at the same time or go together in some other way, even the simplest of brains can learn the association. Moreover, cueing of relationships can produce what is called conditioned learning. We all have heard about Pavlov’s dogs. But even animals as primitive as flatworms can exhibit conditioned learning. If worms are shown flashes of light, not much happens. If they are given mild electrical shocks to the body, the body contracts. If then a flash of light is delivered just prior to an electrical shock, after enough repetitions, the worm starts contracting when it first detects the light, before any electrical shock is delivered.
Associations are still more powerful when they are consciously constructed. This is the stage where you ask yourself such questions as: Where does this information fit with what I already know? How does this relate to other things I could learn about? What value do I place on this information? How invested in using or remembering it should I be?
Then there is understanding. You can, as I did, pass college calculus by using the right formulas for given problem types, and yet not really understand what is going on with the equations. To understand, you need to answer such questions as: Is this consistent with what I thought I knew? What is missing or still confusing? What can I do with this information? What else does it appoly to, how can it be extended? What is predictable?
Learning is not complete without understanding. Understanding also creates a basis for generate insights and creative syntheses, and these in turn advance the depth and rigor of the original learning. Insights typically come from deduction or induction. Deduction is the Sherlock Holmes process of using one fact or observation to lead logically to another. Induction is the Charles Darwin process of using multiple, apparently unrelated, facts or observations to make a synthesis that accommodates them all.
Finally, there is learning to learn. This is the process of learning the paradigm, the “rules of the game,” that allows you to transfer one learned capability to new learning situations that are related. At this point, one has reached a threshold where the more you know, the more you can know.
One of the first experimental demonstrations of this phenomenon was by H. C. Blodgett in 1929. He studied maze behavior in rats, scoring how many errors they made in running the maze to find the location where a food reward was placed. Rats ran the maze once per day on successive days. The control group ran the maze and found the food, with number of errors decreasing slowly on successive days as they learned where in the maze the food was. Experimental groups ran the maze daily for three or seven days without any food reward. Naturally, they made many errors because there was nothing to learn. However, when they subsequently were allowed access to a food reward, the number of errors dropped precipitously on the very next day’s trial. In other words, the rats had been learning about the maze, its layout, number of turns, etc. during the initial explorations when no reward was available.
Blodgett called this “latent learning,” an idea expanded and formalized some 20 years later in the “Learning Set” theory of Harry Harlow. Harlow studied visual discrimination learning in monkeys and observed that visual and other types of discrimination problems progressed more quickly as a function of training on a series of different, but related problems.
These discoveries were born of necessity, arising from the need to use the same monkeys over and over in a wide variety of experiments because the Harlow lab was so under-funded. Increasing the number of problems on which monkeys were tested led to the observation that the monkeys’ general learning competence improved over time. This of course parallels the general common experience of maturation of children.
Harlow developed the prominent theory that learning any task is associated with implicit learning capabilities that can generalize to other related learning situations. The concept relates simpler trial-and-error learning to more advanced insightful-like learning, which he regarded as a mental ability that depended heavily on prior learning sets. Ability to form learning sets varies with species. Monkeys do it better than dogs or cats, and humans do it best of all.The reasons for human superiority in learning no doubt include the rich connections among various brain areas that can support and integrate more learned associations.
Learning involves at least four major processes. It all begins with registering new information. This is the stage when information is detected and encoded in brain. Paying attention obviously facilitates the registration process. Multi-tasking can create an information overload in which much of the information never gets registered. Example: a car driver who is all wrapped up in a cell phone conversation may not realize she just ran a stop sign or cut off the driver behind her in the next lane. Another example comes with reading. Reading comprehension (learning) depends heavily on the eyes actually seeing each cluster of words. The reader needs to focus on words, not letters, and needs to think about what the words mean. Likewise in images, what you learn from an image depends on the details in it that you actually notice and think about.
Next is integration. The brain likes to classify, categorize, and organize its information. Thus, new information has to be fitted into existing learned schema. This is the stage where associations are made with existing memory. Brains are really good at detecting and constructing relationships. If a given relationship is not immediately obvious, the brain may figure it out and remember it. Constructing such relationships is an integral part of the learning process.
Associations can be constructed subconsciously. If two things happen at the same time or go together in some other way, even the simplest of brains can learn the association. Moreover, cueing of relationships can produce what is called conditioned learning. We all have heard about Pavlov’s dogs. But even animals as primitive as flatworms can exhibit conditioned learning. If worms are shown flashes of light, not much happens. If they are given mild electrical shocks to the body, the body contracts. If then a flash of light is delivered just prior to an electrical shock, after enough repetitions, the worm starts contracting when it first detects the light, before any electrical shock is delivered.
Associations are still more powerful when they are consciously constructed. This is the stage where you ask yourself such questions as: Where does this information fit with what I already know? How does this relate to other things I could learn about? What value do I place on this information? How invested in using or remembering it should I be?
Then there is understanding. You can, as I did, pass college calculus by using the right formulas for given problem types, and yet not really understand what is going on with the equations. To understand, you need to answer such questions as: Is this consistent with what I thought I knew? What is missing or still confusing? What can I do with this information? What else does it appoly to, how can it be extended? What is predictable?
Learning is not complete without understanding. Understanding also creates a basis for generate insights and creative syntheses, and these in turn advance the depth and rigor of the original learning. Insights typically come from deduction or induction. Deduction is the Sherlock Holmes process of using one fact or observation to lead logically to another. Induction is the Charles Darwin process of using multiple, apparently unrelated, facts or observations to make a synthesis that accommodates them all.
Finally, there is learning to learn. This is the process of learning the paradigm, the “rules of the game,” that allows you to transfer one learned capability to new learning situations that are related. At this point, one has reached a threshold where the more you know, the more you can know.
One of the first experimental demonstrations of this phenomenon was by H. C. Blodgett in 1929. He studied maze behavior in rats, scoring how many errors they made in running the maze to find the location where a food reward was placed. Rats ran the maze once per day on successive days. The control group ran the maze and found the food, with number of errors decreasing slowly on successive days as they learned where in the maze the food was. Experimental groups ran the maze daily for three or seven days without any food reward. Naturally, they made many errors because there was nothing to learn. However, when they subsequently were allowed access to a food reward, the number of errors dropped precipitously on the very next day’s trial. In other words, the rats had been learning about the maze, its layout, number of turns, etc. during the initial explorations when no reward was available.
Blodgett called this “latent learning,” an idea expanded and formalized some 20 years later in the “Learning Set” theory of Harry Harlow. Harlow studied visual discrimination learning in monkeys and observed that visual and other types of discrimination problems progressed more quickly as a function of training on a series of different, but related problems.
These discoveries were born of necessity, arising from the need to use the same monkeys over and over in a wide variety of experiments because the Harlow lab was so under-funded. Increasing the number of problems on which monkeys were tested led to the observation that the monkeys’ general learning competence improved over time. This of course parallels the general common experience of maturation of children.
Harlow developed the prominent theory that learning any task is associated with implicit learning capabilities that can generalize to other related learning situations. The concept relates simpler trial-and-error learning to more advanced insightful-like learning, which he regarded as a mental ability that depended heavily on prior learning sets. Ability to form learning sets varies with species. Monkeys do it better than dogs or cats, and humans do it best of all.The reasons for human superiority in learning no doubt include the rich connections among various brain areas that can support and integrate more learned associations.
Thursday, December 4, 2008
Make Them Learn: With Carrot or Stick?
Feedback is essential for learning. Not only does the feedback need to ensure that learning was achieved (as in testing), but feedback also needs to reinforce the motivation to learn. The age-old questions arises: do we use the carrot or the stick? Which works best, negative or positive reinforcement? Most people have an opinion, but now we have scientific studies of the question. And the answer is, it depends.
For example, three studies showed that adults correct their behavior better in response to negative feedback rather than positive feedback, whereas 8- to 11-year old children respond just the opposite. A follow-up study by Anna van Duijvenvoorde and colleagues in the
Regardless of the nature of the feedback, young adults learned bdetter than the children. For all three age groups, learning was more effective with positive feedback. Moreover, the decreased learning from negative feedback was conspicuously greater in the youngest age group, while in the young adults, the effect of feedback type was negligible.
Not surprisingly, there were brain scan indicators of differing response to type of feedback. With age, both types of feedback produced a shift toward recruiting more activity in the dorsolateral prefrontal cortex. This part of the cortex was more active after negative feed back in adults but after positive feedback in the 8-9 year-old children. The prefrontal cortex activity was about the same for negative and positive feedback in the 11-13 year olds, suggesting that this is a transition stage in development of learning style and capability.
Take home message? Positive feedback usually works best in young children (that is, after all, how they train seals). Negative feedback works just about as well as positive feedback in young adults. One more point: with the exception of language acquisition, young children are not the superior learners that many people believe.
Source: van Duijvenvoorde, A. C. K. 2008. Evaluating the negative or valuing the positive? Neural mechanisms supporting feedback-based learning across development. J. Neuroscience. 28 (38) 9495-9503.
Subscribe to:
Posts (Atom)