Posts tagged learning
Articles and news from the latest research reports.
How brains remember and correct
Information processing in the brain is complex and involves both the processing of sensory inputs and the conversion of those inputs into behavior. The passing of electrical oscillations between networks of neurons in different parts of the brain is thought to be a critical component of cognition as well as conscious perception and awareness, but so far there has been little direct evidence linking specific neuronal oscillations to discrete thinking and behavior events.
Jun Yamamoto and colleagues from the RIKEN–MIT Center for Neural Circuit Genetics have now detected a brief burst of nerve activity oscillating in two specific parts of the mouse brain just before a correct choice is made, either when planning an action or when correcting a mistake.
The researchers searched for evidence of specific neuronal oscillations by studying mice navigating a T-shaped maze with a reward at the end of one arm of the T. Just before trained mice made the correct choice of direction, Yamamoto and his colleagues observed a brief burst of synchronized high-frequency gamma waves oscillating in specific parts of the entorhinal cortex and hippocampus.
Yamamoto was fascinated to notice that the burst of gamma waves also occurred just before mice that had originally turned in the wrong direction realized their mistake and turned round. He called this the “oops” moment, and the results indicate that similar neuronal activity occurs when making a correct choice either immediately or on realization of an error. No such gamma-wave activity was detected when mice made the wrong choice without correcting it.
To further test the link between the gamma synchrony and the memory recall process, the researchers genetically engineered mice with light-activated ion channels that could block the gamma waves. When these channels were activated, the gamma waves ceased and the mice could no longer accurately choose the right direction or correct their wrong choices.
“Our work is telling us about how the brain recalls remembered information at critical moments,” says Yamamoto. “It suggests that synchronized gamma oscillations actually contribute to the animal’s correct choice rather than being a consequence of their choice.” The finding sheds light on the fundamental mechanism underlying the successful retrieval of working memory. Yamamoto now intends to see if these initial findings apply to other brain regions.
The results also provide new insight into the phenomenon of animal consciousness. “Our findings provide evidence that animals employ a behavior monitoring process called metacognition that typically requires conscious awareness,” says Yamamoto.
Releasing the brakes for learning
Learning can only occur if certain neuronal “brakes” are released. As the group led by Andreas Lüthi at the Friedrich Miescher Institute for Biomedical Research has now discovered, learning processes in the brain are dynamically regulated by various types of interneurons. The new connections essential for learning can only be established if inhibitory inputs from interneurons are reduced at the right moment. These findings have now been published in Nature.
Image caption: Example of a dendrite of a principal neuron (white) and synaptic contacts (yellow arrowheads) from SOM1 interneurons.
For some years, most neurobiologists studying learning processes have assumed that the new connections required for learning can only be established and ultimately reinforced if certain neuronal “brakes” are released – a process known as disinhibition. It has also been supposed for some time that various types of interneurons could be involved in disinhibition. Interneurons are nerve cells that surround and – via their connections – inhibit the activity of principal neurons. It has not been clear, however, whether these cell types actually play a role in disinhibition and how they control learning.
Andreas Lüthi and his group at the Friedrich Miescher Institute for Biomedical Research have now demonstrated for the first time how a learning process is dynamically regulated by specific types of interneurons.
In Lüthi’s experiments, mice were trained to associate a sound with an unpleasant stimulus, so that the animals subsequently knew what would happen when they heard the auditory cue. The researchers showed that, during the learning process, the sound stimulus released a brake in some of the principal neurons. More precisely, it induced the activation of parvalbumin-positive (PV+) interneurons, leading indirectly – via somatostatin-positive (SOM+) interneurons – to disinhibition of the principal neurons. The latter thus became receptive to further sensory inputs. If this was immediately followed by the unpleasant stimulus, then another brake was released. Once again, PV+ interneurons were involved, but this time the principal neurons were directly disinhibited. Steffen Wolff, a postdoc in Lüthi’s group and first author of the publication, explains: “The principal neurons temporarily reached a level of activation enabling neuronal connections to be reinforced in such a way that the animal could learn the association between the sound and the unpleasant stimulus.”
Lüthi comments: “This is the first time we’ve been able to identify so clearly the function of defined interneurons in a learning process, and to show how successive disinhibition can enable this process. We assume that interneurons disinhibit the principal neurons in a highly dynamic manner. They integrate, as it were, the state of numerous different neural networks, activated for example by sensory input, earlier experiences or emotional states, and thus permit or prevent learning. I think these findings are also of interest in the context of conditions where learning processes are impaired or dysfunctional, as in the case of anxiety disorders.”
(Source: fmi.ch)
Neuroscientists discover adaptation mechanisms of the brain when perceiving letters of the alphabet
The headlights – two eyes, the radiator cowling – a smiling mouth: This is how our brain sometimes creates a face out of a car front. The same happens with other objects: in house facades, trees or stones – a “human face” can often be detected as well. Prof. Dr. Gyula Kovács from Friedrich Schiller University Jena (Germany) knows the reason why. “Faces are of tremendous importance for human beings,” the neuroscientist explains. That’s why in the course of the evolution our visual perception has specialized in the recognition of faces in particular. “This sometimes even goes as far as us recognizing faces when there are none at all.”
Until now the researchers assumed that this phenomenon is an exception that can only be applied to faces. But, as Prof. Kovács and his colleague Mareike Grotheer were able to point out in a new study: these distinct adaptation mechanisms are not only restricted to the perception of faces. In the The Journal of Neuroscience the Jena researchers have proved that the effect can also occur in the perception of letters.
(Figure 1: Correspondence between the activity of the medial prefrontal cortex and study results in the second year versus the first year. Horizontal axis shows the degree of activity in the medial prefrontal cortex of various students; vertical axis shows performance improvement in the second academic year compared with the first.)
How the brain builds on prior knowledge
It is easier to learn something new if you can link it to something you already know. A specific part of the brain appears to be involved in this process: the medial prefrontal cortex. The Journal of Cognitive Neuroscience has published these findings, from research by neuroscientists at Radboud university medical center and Radboud University, as an Early Access paper. The findings further enhance our understanding of the brain mechanisms that underlie effective learning.
Neuroscientist Marlieke van Kesteren tested two groups of students who had just started on their second-year of biology or pedagogy studies. While an MRI scanner was registering their brain activity, the students learned short sentences containing new information that expanded on their own or the other study programme. The following day, the students were tested on the information they had learned. As expected, they had retained the information that was related to their own programme better than the unrelated information.
In practice
During the successful retention of related information, a different part of the brain was active than when unrelated information was memorised. ‘The brain area we found, the medial prefrontal cortex, probably linked new information directly to prior knowledge’, Van Kesteren said. ‘In previous studies this brain area came to the fore as well, but only during simple tests. We have specifically shown that this area also plays a role in the neural basis of learning in educational practice.’
Link to study results
To her amazement, Van Kesteren also discovered that the activity in the medial prefrontal cortex corresponded with how well students performed in their second year, compared with the first. So is it possible to predict a student’s future academic success by placing him or her in a scanner? ‘No, certainly not, the links we found were not strong enough’, Van Kesteren explained. ‘We’re mostly talking here about differences of not more than 10% (Figure 1). What’s more, we can’t tell from a simple correlation like this what the chief reason is, and whether a whole lot of other factors are playing a role. But if we know exactly how our brain uses prior knowledge, we could try to address that knowledge more selectively before we start learning new information. For example, you could consider how the new information is related to what you already know.’
Van Kesteren added a tip for secondary school students taking their final exams: ‘If you don’t immediately know the answer to a question, you could first try recalling what you already know about that topic. This might help you to come up with the right answer after all.’
This publication is part of Marlieke van Kesteren’s PhD research, for which she obtained her doctorate at Radboud University Nijmegen in March 2013. In April 2013 she received a Rubicon grant from the Netherlands Organisation for Scientific Research (NWO), allowing her to work on her research into prior knowledge and memory at Stanford University in California for the next two years.
(Image caption: Dendrite of an amygdala principal neuron with dendritic spines (white). Inhibitory synaptic contacts are shown in red. Credit: © MPI f. Brain Research/ J. Letzkus)
A brain capable of learning is important for survival: only those who learn can endure in the natural world. When it learns, the brain stores new information by changing the strength of the junctions that connect its nerve cells. This process is referred to as synaptic plasticity. Scientists at the Max-Planck Institute for Brain Research in Frankfurt, working with researchers from Basel, have demonstrated for the first time that inhibitory neurons need to be at least partly blocked during learning. This disinhibition is a bit like taking the foot off the brake in a car: if the inhibitory neurons are less active, learning is accelerated.
Learning is often a matter of timing: different stimuli become strongly associated if they occur in close succession. The Max Planck scientists made use of this phenomenon in conditioning experiments in which mice learned to react to a tone. For this learning effect to occur, the synapses of the so-called principal neurons in the amygdala need to become more sensitive. The researchers concentrated on two types of inhibitory neurons which produce the proteins parvalbumin and somatostatin and inhibit the principal neurons of the amygdala.
The results obtained by the Max Planck researchers show that both cell types are inhibited during different phases of the learning process. This disinhibition enhances the activation of the principal neurons. Moreover, the scientists were able to control the learning behaviour of the mice through the use of optogenetics. In these experiments, they equipped both types of inhibitory neurons in the amygdala with light-sensitive ion channels, allowing them to use light to switch the neurons on or off as required. “When we prevent disinhibition, the mice learn less well. In contrast, enhancing the disinhibition leads to intensified learning”, says Johannes Letzkus from the Max Planck Institute for Brain Research. Next, the scientists aim to identify the nerve pathways which are involved in disinhibition.
Motor cortex shown to play active role in learning movement patterns
Skilled motor movements of the sort tennis players employ while serving a tennis ball or pianists use in playing a concerto, require precise interactions between the motor cortex and the rest of the brain. Neuroscientists had long assumed that the motor cortex functioned something like a piano keyboard.
"Every time you wanted to hear a specific note, there was a specific key to press," says Andrew Peters, a neurobiologist at UC San Diego’s Center for Neural Circuits and Behavior. "In other words, every specific movement of a muscle required the activation of specific cells in the motor cortex because the main job of the motor cortex was thought to be to listen to the rest of the cortex and press the keys it’s directed to press."
But in a study published in this week’s advance online publication of the journal Nature, Peters, the first author of the paper, and his colleagues found that the motor cortex itself plays an active role in learning new motor movements. In a series of experiments using mice, the researchers showed in detail how those movements are learned over time.
"Our finding that the relationship between body movements and the activity of the part of the cortex closest to the muscles is profoundly plastic and shaped by learning provides a better picture of this process," says Takaki Komiyama, an assistant professor of biology at UC San Diego who headed the research team. "That’s important, because elucidating brain plasticity during learning could lead to new avenues for treating learning and movement disorders, including Parkinson’s disease."
With Simon Chen, another UC San Diego neurobiologist, the researchers monitored the activity of neurons in the motor cortex over a period of two weeks while mice learned to press a lever in a specific way with their front limbs to receive a reward.
"What we saw was that during learning, different patterns of activity—which cells are active, when they’re active—were evident in the motor cortex," says Peters. "This ends up translating to different patterns of activity even for similar movements. Once the animal has learned the movement, similar movements are then accompanied by consistent activity. This consistent activity moreover is totally new to the animal: it wasn’t used early in learning even with movements that were similar to the later movement."
"Early on," Peters says, "the animals will occasionally make movements that look like the expert movements they make after learning. The patterns of brain activity that accompany those similar early and late movements are actually completely different though. Over the course of learning, the animal generates a whole new set of activity in the motor cortex to make that movement. In the piano keyboard analogy, that’s like using one key to make a note early on, but a different key to make the same note later."
Investigators Discover How Key Protein Enhances Memory and Learning
Case Western Reserve researchers have discovered that a protein previously implicated in disease plays such a positive role in learning and memory that it may someday contribute to cures of cognitive impairments. The findings regarding the potential virtues of fatty acid binding protein 5 (FABP5) — usually associated with cancer and psoriasis — appear in the May 2 edition of The Journal of Biological Chemistry.
“Overall, our data show that FABP5 enhances cognitive function and that FABP5 deficiency impairs learning and memory functions in the brain hippocampus region,” said senior author Noa Noy, PhD, a professor of pharmacology at the School of Medicine. “We believe if we could find a way to upregulate the expression of FABP5 in the brain, we might have a therapeutic handle on cognitive dysfunction or memory impairment in some human diseases.”
FABP5 resides in many tissues and is especially highly expressed in the brain. Noy and her Case Western Reserve School of Medicine and National Institute on Alcohol Abuse and Alcoholism colleagues particularly wanted to understand how this protein functioned in neurons. They performed imaging studies comparing the activation of a key transcription factor in the brain tissue of normal mice and in FABP5-deficient mice. (Transcription factor is a protein the controls the flow of genetic information). The investigations revealed that FABP5 performs two different functions in neurons. First, it facilitates the degradation of endocannabinoids, which are neurological modulators controlling appetite, pain sensation, mood and memory. Second, FABP5 regulates gene expression, a process that essentially gives cells their marching orders on structure, appearance and function.
“FABP5 improves learning and memory both because it delivers endocannabinoids to cellular machinery that breaks them down and because it shuttles compounds to a transcription factor that increases the expression of cognition-associated genes,” Noy said.
Even though endocannabinoids affect essential physiological processes from appetite to memory, the “cannabinoid” part of the word signifies that these natural biological compounds act similarly to drugs such as marijuana and hashish. Too much endocannabinoid can lead to impaired learning and memory.
In simple terms, FABP5 transports endocannabinoids for processing. FABP5 functions like a bus and carries the brain’s endocannabinoids and their biological products to two stations within the neuron cell. FABP5 captures endocannabinoids entering the neuron and delivers them to an enzyme that degrades them (station 1). Then, that degraded product is picked up by the same protein (FABP5) and shuttled to the cell nucleus — specifically, to a transcription factor within it (station 2). Binding of the degraded product activates the transcription factor and allows it to induce expression of multiple genes. The genes that are induced in this case tell the cells to take steps that promote learning and memory.
Noy and associates also compared memory and learning in FABP5-deficient mice and in normal ones. In one test, both sets of mice repeatedly swam in mazes that had a platform in one established location where they could climb out of the water. During subsequent swims, the wild-type mice reached the platform quickly because they had learned — and remembered — its location. Their FABP5-deficient counterparts took much longer, typically finding the platform’s location by chance.
“In addition to regulating cell growth as in skin and in cancer cells, for example, FABP5 also plays a key role in neurons of the brain,” Noy said. “FABP5 controls the biological actions of small compounds that affect memory and learning and that activate a transcription factor, which regulates neuronal function.”
(Source: casemed.case.edu)
Preparing for adulthood: thousands upon thousands of new cells are born in the hippocampus during puberty, and most survive with effortful learning
The dentate gyrus of the hippocampal formation generates new granule neurons throughout life. The number of neurons produced each day is inversely related to age, with thousands more produced during puberty than during adulthood, and many fewer produced during senescence. In adulthood, approximately half of these cells undergo apoptosis shortly after they are generated. Most of these cells can be rescued from death by effortful and successful learning experiences (Gould et al., 1999; Waddell and Shors, 2008; Curlik and Shors, 2011). Once rescued, the newly-generated cells differentiate into neurons, and remain in the hippocampus for at least several months (Leuner et al., 2004). Here, we report that many new hippocampal cells also undergo cell death during puberty. Because the juvenile brain is more plastic than during adulthood, and because many experiences are new, we hypothesized that a great number of cells would be rescued by learning during puberty. Indeed, adolescent rats that successfully acquired the trace eyeblink response retained thousands more cells than animals that were not trained, and those that failed to learn. Because the hippocampus generates thousands more cells during puberty than during adulthood, these results support the idea that the adolescent brain is especially responsive to learning. This enhanced response can have significant consequences for the functional integrity of the hippocampus. Such a massive increase in cell proliferation is likely an adaptive response as the young animal must emerge from the care of its mother to face the dangers, challenges, and opportunities of adulthood.
Scientists Identify Critical New Protein Complex Involved in Learning and Memory
Scientists from the Florida campus of The Scripps Research Institute (TSRI) have identified a protein complex that plays a critical but previously unknown role in learning and memory formation.
The study, which showed a novel role for a protein known as RGS7, was published April 22, 2014 in the journal eLife, a publisher supported by the Howard Hughes Medical Institute, the Max Planck Society and the Wellcome Trust.
“This is a critical building block that regulates a fundamental process—memory,” said Kirill Martemyanov, a TSRI associate professor who led the study. “Now that we know about this important new player, it offers a unique therapeutic window if we can find a way to enhance its function.”
The team looked at RGS7 in the hippocampus, a small part of the brain that helps turn short-term memory in long-term memory.
The scientists found the RGS7 protein works in concert with another protein, R7BP, to regulate a key signaling cascade that is increasingly seen as a critical to cognitive development. The cascade involves the neurotransmitter GABA, which binds to the GABAb receptor and opens inhibitory channels known as GIRKs in the cell membrane. This process ultimately makes it more difficult for a nerve cell to fire.
This process turned out to be critical to normal functioning, as the research showed mice lacking RGS7 exhibited deficits in learning and memory.
Martemyanov believes the findings could ultimately have broad therapeutic application. “GIRK channels are implicated in a range of neuropsychiatric conditions, including drug addiction and Down’s syndrome, that result from a disproportionate increase in neuronal inhibition as a result of greater mobilization of these channels,” he said. “Now that we know the identity of the critical modulator of GIRK channels we can try to find a way to increase its power with the hopes of reducing the inhibitory overdrive, and that might potentially alleviate some of the disruptions seen in Down’s syndrome. It is possible that similar strategies might apply for dealing with addiction, where adaptations in the GABAb-GIRK pathway play a significant role.”
Targeting the RGS7 protein could allow for better therapeutic outcomes with fewer side effects because it allows for fine tuning of the signaling, according to Olga Ostrovskaya, the first author of the study and a member of Martemyanov’s lab, who sees many ways to follow up on the findings.
“We’re looking into how RGS7 is involved in neural circuitry and functions tied to the striatum, another part of the brain responsible for procedural memory, mood disorders, motivation and addiction,” Ostrovskaya said. “We may uncover the RGS7 regulation of other signaling complexes that may be very different from those in hippocampus.”
(Image caption: During the learning processes, extensions grow on neurons. Synapses are located at the end of these extensions (left: as seen in nature; right: reconstruction). When the synapse growth is based on the correlated development of all synaptic components, it can remain stable for long periods of time. Credit: © MPI of Neurobiology/ Meyer)
Synapses – stability in transformation
Nothing lasts forever. This principle also applies to the proteins that make up the points of contact between our neurons. It is due to these proteins that the information arriving at a synapse can be transmitted and then received by the next neuron. When we learn something, new synapses are created or existing ones are strengthened. To enable us to retain long-term memories, synapses must remain stable for long periods of time, up to an entire lifetime. Researchers at the Max Planck Institute of Neurobiology in Martinsried near Munich have found an explanation as to how a synapse achieves remaining stable for a long time despite the fact that its proteins must be renewed regularly.
Learning in the laboratory
“We were interested first of all in what happens to the different components of a synapse when it grows during a learning process,” explains study leader Volker Scheuss. An understanding of how the components grow could also provide information about the long-term stability of synapses. Hence, the researchers studied the growth of synapses in tissue culture dishes following exposure to a (learning) stimulus. To do this, they deliberately activated individual synapses using the neurotransmitter glutamate: scientists have long known that glutamate plays an important role in learning processes and stimulates the growth of synapses. Over the following hours, the researchers observed the stimulated synapses and control synapses under a 2-photon microscope. To confirm the observed effects, they then examined individual synapses with the help of an electron microscope. “When you consider that individual synapses are only around one thousandth of a millimetre in size, this was quite a Sisyphean task,” says Tobias Bonhoeffer, the Director of the department where the research was carried out.
Synaptic stability – a concerted effort
The scientists discovered that during synapse growth the different protein structures always grew coordinated with each other. If one structural component was enlarged alone, or in a way that was not correctly correlated with the other components, its structural change would collapse soon after. Synapses with such incomplete changes cannot store any long-term memories.
The study findings show that the order and interaction between synaptic components is finely tuned and correlated. “In a system of this kind, it should be entirely possible to replace individual proteins while the rest of the structure maintains its integrity,” says Scheuss. However, if an entire group of components breaks away, the synapse is destabilised. This is also an important process given that the brain could not function correctly without the capacity to forget things. Hence, the study’s results provide not only important insight into the functioning and structure of synapses, they also establish a basis for a better understanding of memory loss, for example in the case of degenerative brain diseases.