"My aim is to put the pieces of the puzzle together"
From mice to humans, understanding how brains learn and adapt requires connecting many pieces of evidence. Ulman Lindenberger explains how animal models advance research on brain plasticity and healthy aging.
When you are asked as a researcher of aging about the efficacy of brain training, your response is often sobering. Why?
In the field of cognitive aging research, hopes were high for a long time: If the brain is malleable—that is, plastic—, it should be possible to train cognitive abilities in a targeted manner and improve them permanently. That is why decades of research went into developing training programs—for perceptual speed or working memory, for example—with the expectation that a few weeks of regular practice would make people quicker, more attentive, and even more intelligent in general.
However, counter to these hopes, it was found that such training only has limited effects. One usually gets better at the trained task, sometimes with transfer to tasks that are very similar. But it is very rare to find that entire ability domains improve, and if so, the effects do not persist over time. The bottom line is that training improves specific skills rather than abilities. This disenchantment has opened up some basic questions: What do we exactly mean when we talk about "plasticity"—that is, the malleability of the brain? And which changes can really be initiated by experience—particularly in adulthood?
At this point we were confronted with an important decision: Either we continue as before or we take a step back and try to understand the underlying mechanisms systematically. We decided for the latter.
Going back to the basics led to the expansion-exploration-selection-refinement theory proposed by Martin Lövdén and you. Could you explain it?
The origin of experience-dependent plasticity is a mismatch between the demands of a task and what a brain can perform at a certain moment. If existing routines and networks are not sufficient, the brain initially reacts with an expansion: It makes more "material" available—usually by forming additional synaptic connections. The space of potential connections grows larger, and in some cases, a local expansion of the specific brain region is observable in humans.
In the ensuing exploration phase, the brain tries out different neuronal solutions within this extended space. We assume that this exploration occurs via mechanisms of reinforcement learning. Solutions that work well are rewarded and thereby strengthened, whereas less successful ones lose importance. In this phase, the different solutions are rivals to each other. In the end, the most efficient one prevails—that is the phase of selection.
As soon as this solution has been found, the initially established, now superfluous connections are removed. This process is called pruning. It prevents the brain from getting ever larger with each new skill attained. Thereafter follows refinement, during which the selected solution is further optimized and stabilized. In animal models, we can clearly see these phases: expansion, exploration, selection, and refinement. We assume that they also exist in humans.
And that is why you are trying to bridge the gap between animal and human research?
Yes, because many of these processes cannot be observed in humans directly. In animal models—for instance, in mice—we can measure synaptic changes in great detail. We see the typical phases of plasticity there, a short-term overproduction of connections and later pruning, during motor learning, for example.
Our approach now consists in connecting well understood findings on plasticity in animal models with learning tasks in humans that are as similar to the animal model as possible. In this way, we can better interpret the indirect measures taken in humans.
Which methods do you use to test your model?
In humans, we do not have a method allowing us to observe synaptic growth directly. We can use non-invasive methods such as structural and functional magnetic resonance imaging (MRI), positron emission tomography (PET), electroencephalography (EEG), and magnetic resonance spectroscopy (MRS). The evidence we obtain is thus indirect: We assess changes in volume, patterns of activity, and metabolites.
MRS is important in this context because it can detect metabolites. Their pattern indicates whether the so-called excitatory-inhibitory balance, that is, the equilibrium between neuronal excitation and inhibition, moves in the direction of more excitation for a certain period. This shift is known to facilitate the formation of new synapses.
A second element is the layer-specific analysis of the cortex. Using an MRI scanner with high field strength, such as a 7-Tesla scanner, signals can be separated by layer. This makes it possible to identify the layers of the cortex in which volume or activity changes occur. The spatial resolution of MRI scanners with lower field strengths is not sufficient for this purpose. I therefore very much hope that the Institute's 7-Tesla scanner can start operations very soon.
You also use PET in collaboration with colleagues from Sweden. Why?
PET works with weakly radioactive ligands, or markers, that bind to certain molecules in the brain. The binding signal allows for the indirect assessment of metabolic processes. Interestingly, some newly developed ligands can help measure synaptic density. Together with colleagues from Sweden, particularly Martin Lövdén from the University of Gothenburg, we are currently examining whether synaptic density already increases measurably after a short intensive phase of learning—about one day of demanding motor spatial training. If this works, it would be, to my knowledge, the first demonstration of learning-related changes in synaptic density in humans using PET. This study is ongoing, and represents a central element to the empirical backing of our model.
Which other studies are you carrying out based on your approach?
One study, mainly pursued by Sarah Power and Markus Werkle-Bergner, examines memory in early childhood. It deals with interactions between experience-dependent plasticity and brain maturation. We ask ourselves why people hardly remember events that occurred in their first years of life, even though they learnt a great deal during this period. This phenomenon is called infantile amnesia, and it belongs to the great unsolved riddles of memory research. We have carried out a large-scale longitudinal study called ECHO that is intended to determine the relative contribution of three options: Early memory traces are only formed fragmentarily; they are formed completely, but are lost towards the end of early childhood; they continue to exist after the end of early childhood, but are only accessible with great difficulty.
Another study examines motor learning. In this case, we work with very simple, well-controlled tasks, for instance, when mice grasp a seed of millet or humans use chopsticks to pick up a small food item. Such paradigms allow us to investigate learning processes precisely and to compare changes in motor cortex directly between mice and humans. This reduction to elementary skills is important to get at basic mechanisms of plasticity across species.
A third study is closer to everyday life: In a collaborative project, we will compare young people moving from the countryside to the city with adolescent mice experiencing a change of their environment. Thereby, we want to improve our understanding of how environmental changes elicit brain plasticity (see info box).
In yet another large-scale study, which is currently ongoing, we are examining the plasticity of the brain in childhood by investigating the effects of singing in a choir on the neuronal tracts linking auditory and motor areas.
All of these studies approach the plasticity of the brain and behavior from different perspectives. When we decide what to examine next, we strive to achieve a mixture of precision and everyday relevance, from the grasping of an object with chopsticks to the effects of singing in a choir.
How important is collaboration for such a research program?
Collaboration is absolutely essential. Our work reaches out to Sweden, Munich, Dresden, Leipzig, and further afield. For every research question, we collaborate with those colleagues who can investigate them best—and they do not always need to be here in Berlin, even if geographical proximity does often help.
I am particularly pleased about a development within the Max Planck Society: Increasingly, colleagues collaborate across institute and section boundaries. The cognitive neurosciences are anchored in the Human Sciences Section while the relevant animal models come from the Biology and Medicine Section. Basically, the same questions interest us and we can make more headway together.
Finally, what drives you personally?
I am mainly interested in how humans develop across the lifespan as systems. I want to put the pieces of the puzzle together that allow us understand this question. The plasticity model that Martin Lövdén and I developed seems to be compelling and empirically plausible. However, some 80% of it are not yet sufficiently proven. The individual parts have been shown—some in animal models, others in humans—but how these pieces fit together and whether they can really be interpreted as I would like to is unclear to date. In addition to the work that we are already doing with colleagues elsewhere—for example, in Gothenburg or at the MPI for Biological Intelligence in Martinsried—, I very much hope that we will also be able to find significant pieces of the puzzle here in Berlin.
And to return to the beginning of our conversation: If we can manage to understand the basic processes of learning and memory and their age dependency, better perspectives for applied research will open up. I therefore expect training programs aiming to support older people in the competent management of their everyday lives to become more individualized and more effective than is currently the case.
WILDBRAIN—HOW THE ENVIRONMENT SHAPES THE BRAIN
WildBrain will examine how new places, social encounters, and everyday experiences change the brain. The researchers observe genetically identical mice in three very different habitats: from the classic lab to a large semi-natural structure ("Mouse City") to freely living groups. In parallel, they accompany young adults who have recently moved to Berlin from rural areas and examine how their everyday life, geographical orientation, and social contacts develop in their new location. Using modern methods such as video and sensor data, GPS, MRI, as well as cellular measurements in mice, they want to investigate how behavior and the environment interact to shape the structure of the brain. The overarching aim is to gain a better understanding of individuality in developmental trajectories.
Cooperation partners: MPI for Human Development (Center for Lifespan Psychology, Center for Environmental Neuroscience), MPI for Biological Intelligence, TUD Dresden University of Technology, Technical University of Munich.
