• The interest in biological bases of memory is really twofold these days:
• (1) the realization that many memory deficits stem from anatomical and physiological injuries to the brain
• (2) the realization that the psychological process must have a physiological underpinning.
• Historically, the search for the underlying physiological process has focused on a search for the engram.
• There seem to be several equally vague definitions of this term including:
• (1) the memory trace that is presumably present in the brain after something has been learned;
• (2) the set of changes in the nervous system that represent stored memory; and
• (3) the discrete representation in the nervous system of specific ideas, concepts or behaviors.

• Karl Lashley (1929) began the modern search for the underlying neurophysiological mechanisms of learning & memory.
• Lashley sought to provide evidence for the existence of specific, localized, cortical connections, formed during conditioning.
• Lashley's approach was to teach animals to run through a maze without error.
• Then he lesioned various parts of the brain in his search for the 'location' of the memory--he thought that if he lesioned just the right area, the animal would forget the maze memory, but nothing else.
• Much to his disappointment he never found that removing any specific cortical tissue produced any specific memory loss.
• Furthermore, he found that the larger the lesion he created, the greater the memory loss.
• Because of this he formulated the equipotentiality hypothesis in which he suggested that all parts of the cortex have equal potential to produce learning.
• In retrospect, it is possible that by creating rather large lesions--even the smallest of his lesions were fairly large by today's standards (he used a soldering iron)--he destroyed several types of memory cues all in one fell swoop.
• Lashley’s search for the engram epitomizes the brain distribution position.
• On the other side are the brain localization proponents.
• Exemplified by Broca, Wernicke & Milner, these researchers have provided evidence for particular areas areas as specialized for particular cognitive activities.
• So, is memory localized or distributed? We will work to answer this question.

A.) Transfer of Learning

• Research into the transfer of learning has been fruitless but interesting.
• It was motivated by the notion that maybe memories could be transferred by transplanting neurons from one organism to another; or by transplanting DNA and/or RNA from one animal to another.
• McConnell (1962) found initial support for transfer of learning hypotheses, based on studies with planaria.
• He trained them to produce a conditioned response, then spun down the organisms and extracted their RNA, which he injected into other planaria that had not been trained. These planaria then also showed the conditioned response!
• These findings produced much excitement. However, subsequent studies failed to replicate these findings.

• Current thinking is that memory involves anatomical changes in neurons and in their synaptic connections.
• 1.) The Aplysia Studies: synaptic changes
• Kandel and colleagues (1980's) studied learning in the simple nervous system of the sea snail, Aplysia, hoping for insights into more complex learning in more complex organisms.
• They examined classical conditioning of the gill reflex. During habituation of this simple reflex response there is a decrease in responsivity to repeated stimulation, as well as a decrease in neurotransmitter release presynaptically, with a concurrent decrease of the postsynaptic potential.
• During sensitization of the response, the opposite occurs--there is an increase in responsiveness to a previously neutral stimulus, as well as in neurotransmitter release, with a concomitant increase of the postsynaptic potential.
• 2.) Long-Term Potentiation of Synapses
• LTP is a process where electrical stimulation, especially in the hippocampus, results in long-lasting (up to weeks) increases in excitatory potentials in affected neurons--a subsequent increase in responsivity of that neuron.
• On the one hand, the exact relationship between LTP and increased synaptic conduction has not been able to be worked out.
• Some researchers believe that we may not be able to generalize from animal models to humans.
• On the other hand, because of consistent findings with nonhuman primates and other mammals, the concensus seems to be that although convincing proff that LTP is involved in memory cannot be demonstrated, it still seems to be a viable candidate for synaptic changes related to learning and memory (c.f., Martinez & Derrick, 1996).
• 3.) Effects of Enriched Environments
• Rosenzweig (1972) concluded that rats placed in an enriched environment, even late in life, develop many more active synapses in the cortex than rats that remain in cages--especially in dendritic fields in the visual cortex and hippocampus, as well as in activity of acetycholinesterase (AChE) in the brain.
• Both the number and size of synaptic connections increased, as did cell bodies, which were needed to maintain the increased cell growth.
• All of these effects could be produced throughout the life span, although the effects were more permanent and stronger with younger animals.
• These results have been replicated with birds, mice, gerbils, squirrels & monkeys, as well as lower invertebrates & insects!

• Based on studies with animals and humans with memory impairments, several neurotransmitters have been implicated in memory.
• Acethycholine seems to affect memory by increasing arousal levels.
• Epinephrine, dopamine and serotonin act on limbic structures, especially the amygdala.
• GABA, which is an inhibitory neurotransmitter also affects memory by modulating other cognitive activities.
• Rosenzweig (1996) found changes in AchE and other neurochemicals, especially
• (1) increased rates of protein synthesis & in overall amounts of protein in the brain, and
• (2) increased amounts of RNA following learning--particularly if long-term memories are to be formed (so maybe McConnell was in the right ball park!)

• Consolidation refers to the process by which a memory achieves a stable and permanent form. There are 3 ways to approach the concept of consolidation.
• 1.) From a Structural Standpoint: it refers to the transfer of materials from a short term store to a long term store.
• 2.) From a Process Perspective: it refers to the mechanism by which an item is stored for holding and future retrieval.
• 3.) From a Neurobiological Approach: it refers to consolidation as the result of repeated electrical activity in the neural circuits which causes chemical or structural changes in the neurons themselves, leading to the development of new circuits which them represent new "memory".
• Researchers agree that memory storage is not immediate, but instead develops gradually over time.
• This leads to an examination of retrograde amnesia: the inability to remember events from prior to the onset of amnesia; it is usually temporally graded, so that recent memories are most likely to be lost, compared to more remote memories.

• Many animal studies with ECS have been performed to study retrograde amnesia. There is also a large body of data on humans who have received ECS for therapeutic reasons.
• All of these studies provide consistent insight into consolidation.
• (1) An early active avoidance study in which the animals had to shuttle to the "safe" side of a box to avoid or escape shock was performed by Duncan (1949) with rats.
• Duncan examined the retrograde amnesia gradient--the amount of time between learning and ECS application necessary to result in amnesia.
• Using active avoidance procedures he concluded that memory consolidation takes about one hour.
• This study was criticized, however, in that it was suggested as an alternate explanation for animals failing to learn to escape the shock was that animals preferred the shock to the ECS and so learned to not shuttle in order to avoid the ECS!
• (2) Chorover & Schiller (1965) found that their passive avoidance procedure--inhibition of action--in which the animal had to avoid stepping on a shock grid, reduced the gradient to 10 seconds.
• This seemed a better procedure because it eliminated alternate explanations for the animals failing to shuttle.
• Consolidation still took time, but much less so.
• (3) Squire & Spanis (1984) tested mice with a passive avoidance test over a 4-week period. They found that graded retrograde amnesia lasted for as long as three weeks.
• Thus, the conclusion from ECS studies is that consolidation has no fixed time line, and that its time course can vary widely, which is in accord with human studies where ECS has been used for therapeutic reasons.

• Head injury evidence supports the findings that disruption of consolidation occurs--remote memories may return, but memories for immediate events resulting in the trauma do not.
• Retrograde amnesia involves disruption of the consolidation process, rather than being a simple retrieval deficit--patients would have difficulty recalling all past events, not just those closest to the trauma.
• Thus, studies of retrograde amnesia show that consolidation is not an automatic process of fixed duration, but is a dynamic process that continues so long as information is being forgotten, with a reorganization and stabilization of what remains.
• Other factors may influence the nature of retrograde amnesia, such as strength of original learning, time course of normal forgetting, specific structures damaged, and nature and timing of postoperative testing.

A.) Arousal and Consolidation

• Arousal theory assumes that the amount of memory consolidation is related to the amount of neural activity in the brain, and that the amount of neural activity is a function of arousal level.
• If neural activity is low due to low arousal, less consolidation will occur; if arousal is heightened then consolidation should increase.

• This is made very complex by many factors such as the type of task being learned, the type of drug involved, whether the drug was taken before or after, how much was taken, how long before or after, whether the drug facilitate or interferes with memory, how well the task was learned, how arousing the task itself was, etc.
• The general finding is that any drug that increases nervous system arousal also enhances memory performance.
• Thus, benzodiazepines (BZs--including Valium, Xanax), widely prescribed psychoactive drugs which act to calm or sedate people result in a side effect of impaired memory, specifically in terms of encoding deficits.
• Neurotransmitters known to increase arousal, such as acetylcholine and norepinephrine act to enhance memory performance as well; whereas drugs such as GABA that act as CNS inhibitors tend to inhibit memory performance also.
• As far as stimulants go, the general finding has been that these drugs will enhance memory if given immediately after learning, and if given in only moderate quantities. Too much, and memory will be disrupted because attention is affected.
• Finally, to the extent that glucose increases metabolic activity in the brain it can be expected to, and has been shown to enhance memory.

• Most sleep studies, since early modern psychology, have consistently shown that when sleep immediately follows learning, memory improves.
• The best explanation for this finding seems to be that interference effects are reduced--if one is awake after learning, new information is continually coming in, inhibiting consolidation of target information.

• The sleep studies that have been done during REM sleep suggest that because of the paradoxical cortical activation during REM then consolidation should improve.
• In fact, studies done where subjects were wakened and prevented from REM sleep showed disruption of memory performance.

A.) Electrical Tests

• 1. Electroencephalography (EEG)
• EEG is an imaging technique that measures brain function by analyzing the scalp electrical activity generated by brain structures. Information is gathered from surface electrodes, which are placed using a universally accepted pattern. It is a completely noninvasive procedure.
• Local current flows are produced when brain cells (neurons) are activated. However, only electrical activity generated by large populations of neurons concurrently active can be recorded on the head surface.
• Information is for gross activity levels only, as it must travel through layers of fluid, tissue and bone and subject to much distortion.
• The small electrical signals detected by the scalp electrodes are amplified thousands of times, then displayed on paper or stored to computer memory.
• Clinical applications include the diagnosis of epilepsy, sleep disorders, stroke, head trauma, etc. It is currently being widely used to study the brain organization of cognitive processes as well. For this purpose, the most useful application of EEG recording is the ERP technique.
• 2. Evoked Potentials of Event-Related Potentials (EP or ERP)
• Surface electrical changes are extracted from scalp recordings by computer averaging time frames of EEG, time-locked to repeated occurrences of sensory, cognitive, or motor events.
• Spontaneous background EEG fluctuations are averaged out, leaving the event-related brain potentials, reflecting only that activity which is consistently associated with the stimulus, in a time-locked way.
• The ERP thus reflects, with high temporal resolution, the patterns of neuronal activity evoked by a stimulus.
• Characteristic peaks and valleys emerge, but two waves have been studied the most:
• P3 (P300) occurs about 300 ms after stimulus onset, is often the 3rd peak, and is a positive wave;
• N4 (N400)--occurs about 400 ms after a stimulus, is often the 4th valley, and is a negative wave.
• The amplitude and exact temporal occurrence of these two waves varies under VERY many conditions (I.e., with various pathologies or cognitive activities) and is thought to provide insight into brain activity in response to stimulation.
• Caution in interpretation, however, seem prudent, given the very wide application and interpretation assigned to these waveforms.
• Mental operations, proceed over time ranges of msec--ERPs are an ideal methodology for studying the timing aspects of both normal and abnormal cognitive processes.
• ERP data provide less accurate spatial information than PET or fMRI, which lack fine temporal resolution. Thus, ERPs are a natural complement of PET and fMRI to study human cognition--PET and fMRI can localize regions of activation during a mental task, ERPs can help in defining their time course.
• Some consistent findings relative to memory: rote rehearsal of information results in activation of posterior areas of the cortex (parietal areas). Elaborative rehearsal activates more frontal areas.
• Subsequent Memory Effects: record of ERPs at study time can be compared to ERPs at testing time and compared across items that were retrieved, compared to those that failed to be retrieved.
• Consistent findings are that later-remembered items show more positive-going waves at study time!
• B.) Imaging Tests
• These tests sometimes allow us to view function and not just structure.
• 1. Xray - primarily provides a simple film of skull bones and may show large masses. No functional insights. Strictly anatomic information with possibility of defining major lesion pathology.
• 2. Computerized Axial Tomography (CAT) scans:
• Combines X-rays through multiple planes with computer assistance.
• Can analyze static structures with has greater resolution than the X-ray in showing tissue margins and fluid spaces, but cannot show functional activity.
• 3. Magnetic Resonance Imaging (MRI)
• Shows static structures using magnetic images, in varied planes, with greater resolution than the CAT scan. It is non-invasive and can acquire accurate, high resolution anatomic images in 2D and 3D with excellent soft tissue contrast.
• MRI uses a pulse of radio-frequency energy to excite the protons in the region of interest. Because these protons are in a stable magnetic field, they will only absorb a characteristic energy.
• Although MRI is normally a noninvasive technique, contrast agents (radioactive isotopes) can be injected into the blood stream to enhance a region of interest.
• Intrinsic contrast differences caused by oxygen metabolism, which produce different ratios of oxy- and deoxyhemoglobin, are the basis for interest in cognitive studies using MRI, an emerging technology called functional MRI (fMRI).
• MRI can be used to generate additional information through what is termed localized spectroscopy or magnetic resonance spectroscopy (MRS).
• MRS is the only noninvasive technique capable of directly measuring chemicals within the body.
• Rapid-scanning techniques, which allow the acquisition of an MR image in a time frame of 30 to 100 msec, have opened new applications in MRI including studies in human functional brain imaging.
• 4. Positron Emission Tomography (PET) scans:
• PET is a method of imaging physiologic functions--blood flow and metabolism. A short-lived radiopharmaceutical is injected, which contains a radioactive atom that emits positrons.
• As the positrons encounter electrons in the body, they produce high-energy photons (gamma rays) that can be traced by radiation detectors surrounding the body. Using reconstriction algorithms, images of the distribution of radioactivity throughout the body.
• Positron-emitters include common elements, i.e., oxygen, carbon, nitrogen, etc. Glucose labeled with positron-emitting fluorine (18F) is commonly used to measure energy metabolism.
• Has great spatial resolution in the sense that function can be localized. But has poor temporal resolution--takes minutes to accomplish, compared to the seconds it takes for the cognitive activity.
• A PET camera makes tens of thousands of measurements each second. Each measurement looks at a very small volume of tissue, only about 2 mm³.
• These highly precise measurements create a 3D measurement of an entire organ which can be viewed as images or analyzed in detail by computer programs.
• For cognitive function, typically a subtraction method is used in which baseline readings are taken (1) while there is no specific cognitive activity, and (2) during specific cognitive activity.
• The remaining activity is taken as a measure of processing.
• Temporal resolution of PET scans is greatly improved when the technique is combined with ERP data.
• 5. fMRI - functional MRI
• fMRI is based on the increase in blood flow that accompanies neural activity in the brain. Human cortical functions can be observed without the use of contrast enhancing agents.
• A rapidly emerging body of literature documents corresponding findings between fMRI and conventional electrophysiological techniques to localize specific functions of the human brain.
• The main advantages to fMRI as a technique to image brain activity related to a specific task or sensory process include 1) the signal does not require injections of radioactive isotopes, 2) the total scan time required can be very short, and 3) the resolution of the functional image is precise.
• 6. rCBF - regional Cerebral Blood flow studies
• These are PET studies that are enhanced with the use of injections of radioactive isotopes to monitor blood flow during cognitive processes.

• Anatomy is complicated--we KNOW some regions are involved in memory, but there seems to be no single area that is exclusively involved, or that consistently produces a particular type of memory deficit.
• Short-term memory appears to be sensory specific and may 'reside' in specific sensory cortical areas most closely associated with the input. Since many stimuli are multi-sensory, association areas may be involved.
• Long term memories seem to reside in association cortex areas.

• Based on studies of patients with frontal lobe injury, various memory functions have been attributed to the frontal lobes, including metamemory, some aspects of short term memory and temporal aspects of memory.
• Such findings are not surprising given that the frontal lobes are HIGHLY interconnected to other structures, particularly the hippocampus and limbic structures, thalamus and temporal lobes.
• Shimamura (1995) demonstrated no impairment in learning new information when only the frontal lobes are damaged. While there is some free recall deficit, there is no deficit in recognition, suggesting primarily a search strategy deficit.
• He suggests that the apparent functional disruption based on brain damage might be attributed to disruption in inhibitory control, especially with dorsolateral prefrontal lesions.

• For the most part the only clear-cut memory deficit following injury to the parietal lobes is for topographical material, i.e., the inability to recall the spatial arrangement of familiar surroundings and/or previously well-known geographical relationships.

• There is consensus that a lesion in the temporal lobe in the dominant hemisphere for speech produces impaired learning and memory, regardless of the memory task used.
• With damage to the nondominant temporal lobe deficits clearly appear in memory for nonverbal patterns and information.
• Squire & Knowlton (1995) provide evidence that consolidation appears to originate in the medial temporal lobe as far as long-term memory is concerned.
• However, over time, memory appears to be less constrained to that one area and becomes more diffuse.
• Damage to the medial temporal lobe also results in a deficit in spatial memory abilities.
• Within, and highly interconnected to the medial temporal lobe are other structures which appear to be critically important for memory consolidation, especially for declarative memory.
• These include the hippocampus, the entorhinal cortex, the parahippocampal cortex, and the perirhinal cortex.
• Of these, the hippocampus has received the most attention.

• Limbic structures are particularly implicated: the hippocampus, amygdala, mammillary bodies, dorsomedial thalamic nuclei, and the interconnecting pathways.
• More recently, deficits in terms of declarative and nondeclarative knowledge suggests subcortical involvement for declarative memories; however, nondeclarative memories seem to be more clearly stored in the cerebellum.

• When there is bilateral hippocampal damage there generally is a memory deficit. The typical clinical picture, which primarily occurs following encephalitis, is of a gross deficit in recent memory--no consolidation is apparent after damage occurs--some retrograde amnesia, and little impairment of other intellectual functions.
• The hippocampus is currently considered as primarily instrumental in consolidation, especially of declarative memories and spatial localization. Thus, it is not considered a ‘storage’ site, but a processing site for memory.
• It is not implicated in nondeclarative memories, such as motor skills, which possibly involve noncortical areas primarily.

• There are several lines of evidence that the cerebellum is involved in motor skill learning and memory.
• Specific evidence from eye blink conditioning studies shows that there are changes in cerebellar neural circuits with repeated exposure to stimuli.
• These neural changes appear to occur within the cerebellar cortex as well as deep cerebellar nuclei.