This paper reflects the research and thoughts of a student at the time the paper was written for a course at Bryn Mawr College. Like other materials on Serendip, it is not intended to be "authoritative" but rather to help others further develop their own explorations. Web links were active as of the time the paper was posted but are not updated.

Contribute Thoughts | Search Serendip for Other Papers | Serendip Home Page

Biology 202
2001 Third Web Report
On Serendip

Memory, Learning and Our Virtual Brains: A Hebbian Web

Avis Brennan

Clearly, education has an immediate and definite purpose in my own life. As a student, I have been exposed to events that have had a profound impact on the development of my interests, talents, values, or what some may call intelligence. My own approach to problem solving has inspired me to frame the question of the interplay of intelligence, human behavior, learning, and experience in the language of neuroscience. Neuroscience is appealing to me because it offers an explanation for the micro and macro-level processes that operate to create my complex and unique approach to problem solving and response to the world. In this way, the alignment of neuroscience with biology and psychology offer a convincing general plan for the framework of stimulus and response. That is, neural cell development impacts upon external stimulus reception that, in turn, directs cognitive processing and eventually elicits a behavioral response. This course has indicated to me that the neural circuitry involved with such systems of stimulus and response are not linear, but complex and interdependent. What follows is an integration of brain-based theories on internal representations of experience and memory for the purpose of providing perspective, and ultimately improving the conditions for learning and development in a progressive way.

The term intrinsic variability has been drilled into our brains. No doubt, my dendritic spines have stretched, etching a special "intrinsic variability" path extending from deep in my hippocampus to the folds of my neocortex. When I recall this phrase, a furious stream of chemical spills and reactions is ignited. In my brain, this stream passes through neurons that code for other phrases that have become important to me throughout this course, for example, learning, memory, intelligence, and feedback. But like the term implies, this is my own stream. In my final web paper I will guide you through the landmarks of my path. What follows from intrinsic variability is the idea that the brain is plastic, or impressionable. It seems there is a stability-plasticity trade-off where our brains are hardwired to retain function, yet malleable enough to learn, create memories and, in some cases, compensate for functional damage (1). With each moment that passes, a particular exchange with the external world rearranges some piece of neural circuitry, yielding unique and diverse neural patterns for each individual(2).That no brain ever looks the same from one moment to the next makes the job of understanding brain function inherently difficult.

A majority of the web papers from this semester have discussed how certain neurological disorders result in an "abnormality" in neural processing, perception, or behavior (4). A tangent cascade of neural firing leads me to the question: if we are intrinsically unique, what is the distinction between a difference and a disorder? My first inclination is to assert that prescribed boundaries of normal and disordered neurological function are arbitrary. However, a consideration for the relative relationship of stability and plasticity in the brain suggests that "dysfunction" occurs when otherwise stable processes are disrupted. For example, our brains are composed of units and substances that act in a very organized and specific way. Complex information from the outside world is systematically translated into the electrical and chemical language of the nervous system. Motive behavior likewise relies on intricate systems of excitation and inhibition achieved by the controlled release of precise amounts of neurotransmitters and ions in specific locations (1). My classmates have described some of the diverse ways that disruption in these processes results in disordered (i.e. over-inhibited or over-excited) mood, thoughts and behavior as is the case with Depression, Schizophrenia or Parkinson's disease(4). The sources of such disruption may be environmental, genetic, or more likely, a combination of the two.

While vulnerability to morphological changes is classified as "neurological dysfunction" in certain systems, it is regarded as learning and memory in others. In the case of memory, modifications in the connections between pre-synaptic and post-synaptic neurons establish an assembly of connections. The simultaneous activation of neurons that have been connected in this way is the system by which we retrieve memories (1), (2). This process is relevant to the present discussion for several reasons. First, it offers an example of how we are innately unique. That is, memory is the most active example of how our brains are subject to the influence of experience. The transduction of information from the external world, to the perceptual, and eventually memory-related loci of our brains involves an intricate exchange of input and output. This exchange is dependent on many factors, such as the magnitude of the incoming signal, proximal and temporal relationships between incoming stimuli, the existence of an external, supervising agent, and the existence of other memories (2). Donald Hebb put forth the idea events that occur together in the outside world are bound together in the brain's representation of those events, or our memories (3). The simultaneous activation of different neurons, and the web of pre-existing, and potential connections of each neuron involved in the memory, result in complicated web of associations. The notion that our experiences are represented in a web, woven with a consideration for the activity of the outside world and the activity of our endogenous systems for receiving and processing such information, is a potent example of how brain function is not linear, but an extremely complex interaction and convergence of variables. There are infinite combinations, resulting in different schemes for categorizing memories. These differences are evidence for our intrinsic variability.

The variability in the way that afferent information is processed results in variability in the way that information is subsequently consolidated or retrieved. The theory that memory is the result of neurons that connect and act in assembly, offers useful insight into the way we retain and recall information. Memory is commonly held as a mechanism for storing pieces of important information. This view suggests that remembering is the process of flipping through our catalogue of stored information. However, memory organization is dependent upon myriad contextual cues. Indeed, memories consist of associations, not isolated pieces of information that are shelved in some systematic way.

This logic refutes the commonly held conception of the "memory bank"-- a systematic, identifiable storage mechanism for all the various pieces of information impressed upon us. The associative nature of memory implies that the areas involved in this process are diffuse. Karl Lashley's life work provides evidence for this. He conducted a study wherein he trained dogs and selectively lesioned brain sites to determine the underlying site of memory storage. He failed to find one critical site and concluded that memories are typically spread throughout the brain (3). The works of Lashley and Hebb have therefore provided little hope that we can untangle the formula of memory storage and memory recall. However, the body of literature that has immerged in the last fifty years has elucidated some of the properties and mechanisms of memory, learning and the brain. I will begin with an explanation of what we do know, or what we think we know, about memory and the brain. I will go on to discuss how the construct of memory applies to human learning, and ultimately, how memory and learning are addressed in our attempts to educate and engage children.

In 1954, Scoville discovered that the formation of immediate memories is reliant on a subcortical brain structure called the hippocampus (6). Scoville lesioned the basal ganglia (amygdala and hippocampus) of patient HM, attempting to reduce severe epileptic seizures. HM subsequently experienced an inability to retrieve memories within ten years of the surgery and an inability to form new memories(6), (7). The deficits in information storage and retrieval have contributed to our understanding of memory by distinguishing differences in the way we process information. Specifically, HM experienced normal short-term memory. For example, he could immediately recite a list of numbers. However, if HM were told to study a list he would not remember the list and, moreover, he would not remember ever having studied that list after several hours (7). Thus the case of HM is strong evidence that short-term memory (STM) differs from long-term memory (LTM).

The different subsystems of memory have unique functions, and neural mechanisms. STM acts as our working memory, and has the capacity for a few chunks (7 plus or minus 2) of information. The consolidation of this information determines how a memory will be consolidated in LTM (6). So how is it that we remember single experiences or pieces information that have not been repeated? One theory, put forth by McClelland, McNaughton and O'Reilly (1) is that the Medial Temporal Lobe (MTL) memory system regulated the translation of information to LTM, and ultimately, the trade-off of plasticity and stability in the brain. Their explanation is this: the MTL is an intermediary system for holding memory that is very plastic. This plastic system is receptive to new information and is capable of playing this information back to the less plastic cortex, which stores the information for subsequent retrieval (1).

Hebb's idea that memories are stored in complex, interconnected ways based on the exchange of external and internal information is upheld by patient HM and the theory of LTM memory consolidation described above. That is, memory is a process of neural activity that is determined both by the functional, neurological components and the significance or valence of incoming sensory material. Studies involving brain lesions and damage have indicated that morphological changes result in memory formation. Such evidence is testament for the functional, neurological component of memory-related processing. I will now consider the role of learning in memory, and how the significance or valence of the external stimulus may impact memory formation and learning.

We cannot separate learning and memory. These processes are not isolated, linear experiences. Learning is a modification of memory. A system has "learned" when it has successfully changed its response based on information that was received at an earlier time(1). It seems possible to define memory as a process residing in the exchange of information between functional parts of the brain (i.e. the hippocampus and the cortex). It can be argued that learning describes the reactivation of consolidated memories. Memory formation is rooted in the activity at the synapse (2). Modifications of the synapse, or synaptic weight, underlie our ability to store and recall information. However, it is a mistake to confuse intelligence with these processes. Intelligence describes the way with which we execute important tasks, and therefore learning, and mechanisms that aid in the efficiency of our thoughts or actions are part of how intelligence works. One way that we more efficiently enhance memory-related learning by creating associations that are interesting or emotional in order to facilitate LTM encoding in the cortex. Study techniques are often designed to help an individual create a relevant and personally significant context for the way pieces of information fit together. In this sense, a mnemonic device imposes Hebb's idea of association to result in the desired sequence of encoding among neurons, and ultimately, a predictable sequence for retrieval.

How does such a perspective influence our approach to education? It appears that there are fundamental differences in the way that we learn. There are book-smart people and street-smart people. There are visual-learners and verbal-learners. Labeling students as disabled, gifted or average over generalizes an approach to learning. The kinds of learning for which each brain is specialized require a differential approach to teaching. Each approach should target the form of representation most suitable to the needs of the student. For example, a student with a proclivity toward art may process information more efficiently when it is provided in a visual context. This idea implies that there is a pattern which may underlie the laws of association that Hebb proposed. This pattern may extend to represent the emotionally significant coding of information into LTM. CAST, an organization composed of education and neuroscience researchers who discuss a universal design for learning, propose that teaching recognition should identify and make explicit the rules or patterns to which students gravitate (5). Teachers would present material and offer practice in a context sensitive to the particular pattern of the student in order to achieve the greatest level of engagement and cognitive efficiency. This theory is aligned with the Theory of Multiple Intelligence, proposed in the 1980s by Harvard researcher Howard Gardner. In addition to the language-based logical-mathematical intelligence that is traditionally quantified by measures such as the IQ test, Gardner suggest several other forms of intelligence including spatial, bodily-kinesthetic, musical, intrapersonal or interpersonal (8). In applying this theory to education, Gardner proposes studying the patterns that emerge from compositions, subject and activity choice, and other behavioral and cognitive processes of the student. The patterns may be used to evaluate which intelligence, and thus learning category, is appropriate for that student.

Gardner's theory, while more attuned to the possibility of individual differences, suggests that the variability between brains and within one brain can be segmented into numerable patterns. With this logic, we can redress the problem of disability versus variability. According to the theories described herein, there is a sequential process by which we receive, process and decide to store information. The factors and patterns that determine the precise associations and the mechanisms that determine whether information should be encoded permanently may be different (i.e. verbal versus musical) between people. However, every functional learner is equipped with the capacity to exchange plasticity for stability in the consolidation of LTM to create a network of accessible information. With this in mind, morphological and neurological disturbances in function should not be confused with individual differences in the way information is selected and stored. Which is to say, the deficient transfer of information from the hippocampus to the cortex should not be mistaken for selectivity in information that is received and consolidated.

Learning, memory and intelligence are inexplicably connected. It seems that memory is the activity that underlies the cognitive aspects of learning. Intelligence describes the efficiency with which we learn. Intelligence is therefore in the hands of the media by which we gather and practice pieces of information. Hebb has suggested that we process information through the creation of unique [intrinsic] associations based on experience. Gardner has suggested some of the over-arching contexts that may dictate individuals' information networking. Theories on how memory is translated from STM to LTM suggest that these pattern generations work in tandem such that relevant information is stored in a semi-systematic but non-linear way. Mindfulness for the principles of brain function, such as plasticity and stability, help to demonstrate how these theories fit together to explain differences in learning styles. Traditional education, while rooted in the outmoded methods of linguistic-mathematical assessment, is beginning to incorporate more diverse teaching media to reach varying "kinds" of intelligence. The advent of the World Wide Web has strongly influenced the way we approach learning. This context is designed to appeal to a wide variety of individual "patterns". Moreover, the medium of the Internet represents the advancement of our understanding and appreciation for the importance of networking and interconnectedness in information processing. The Internet is a Hebbian web; a search engine is a virtual cue that represents the unique combination of context and information that results in access to pieces of stored knowledge.

WWW Sources

1) An Academic Resource from Salum , ANS, Structural Changes and Neural Plasticity

2) Serendip Neuroscience Website , May 1, 2001 Lecture notes, by Dr. Grobstein

3) Academic Resource UCL , Memory and Hebb's Rule

5) CAST Homepage , Theory and research involving a universal design for learning

6) Memphis Neuropsychology and Behavioral Neuroscience Homepage , LTM and STM

7) Ohio State Psychology Department Homepage , Patient HM

8) Harvard Graduate School of Education Homepage , Howard Gardner's Theory of Multiple Intelligence

9) University of Newcastle upon Tyne, England Homepage , Evolution and the Cognitive Neuroscience of Awareness, Consciousness and Language, by Bruce Charlton, MD Department of Psychology

| Course Home Page | Forum | Brain and Behavior | Serendip Home |

Send us your comments at Serendip

© by Serendip 1994- - Last Modified: Monday, 07-Jan-2002 14:30:53 EST 1