Navigation and service

3D Model of the Synapses in the Human Brain

For many years, Professor Joachim Lübke has been pursuing the goal of creating high-resolution 3D models of the synapses in the human brain. For a long time, these tiny contacts between neurons could only be studied using animal models. Now, together with his team, Lübke has published the first quantifiable models of synapses in the human cerebral cortex. Their models showed that although similarities exist, there are considerable differences not only between humans and animals but also between men and women.

Prof. Dr. Joachim LübkeProf. Dr. Joachim Lübke
Copyright: Forschungszentrum Jülich / Ralf-Uwe Limbach

“One of the big questions from today’s research into synapses is whether the results that have been and continue to be found in experiments with animals can be applied directly to humans. This is because most of what we know about synapses comes from studies based on various animal models and species,” explains Lübke, who works at Forschungszentrum Jülich’s Institute of Neuroscience and Medicine (INM-10).

Synapses, the contacts through which the neurons communicate with each other, are key elements in the transmission of information. They are so tiny that an optical microscope is not powerful enough to produce a detailed image of them – an electron microscope is needed instead. Their diameter is just a few thousandths of a millimetre on average. Around 100 billion of these junctions can be found in the brain of an adult human. For comparison, if we were to imagine that one synapse was as big as the head of a pin, all of them together would fill 50 long freight trains. Today, many researchers think that this large number of connections holds the secret to the astounding versatility and capability of the human brain.

Perfect shape

The researchers in Lübke’s working group on the structure of synapses are going a step further. They want to capture the structure of individual synapses down to the smallest detail. “Examining high-resolution images of these tiny junctions is a time- and labour-intensive process, but in the end it leads to new, fundamental findings about these structures in both healthy and diseased brains,” says Lübke.

Synapses are always made up of virtually the same components; however, the composition, size, and number of these components has a decisive impact on function. “If we compare synapses from different regions and layers of the brain, or even from different species, we encounter some remarkable differences. It seems likely that synapses do not just form randomly, but instead have a structure that is perfectly adapted to their respective function in a specific network of the brain,” explains Lübke.

Signal transduction at a synapse

Signalübertragung an einer Synapse. Part of the neuron that is emitting a signal ends in a presynaptic element (bouton), which contains several hundred or even several thousand vesicles filled with neurotransmitters. When the signal is transduced, these vesicles merge with the cell membrane and release small precise amounts of the neurotransmitter into the synaptic cleft in a targeted manner. On the other (post-synaptic) side, the released neurotransmitter molecules latch on to specialized neurotransmitter receptors on the receiving neuron, where they are processed and converted, and the resulting signal is transmitted onwards.
Copyright: Forschungszentrum Jülich

Potential approach for new treatments

If there are problems in the transduction of signals in the brain, the consequences can be severe. All neurological and neurodegenerative disorders, such as schizophrenia, autism, Alzheimer’s disease, and Parkinson’s disease, can be attributed to large-scale pathological changes in synapses. What’s more, because of the world’s ageing population, these diseases are becoming increasingly widespread.

New, fundamental findings regarding the structure of synapses could provide an approach for future treatments. But some groundwork still needs to be done. “To understand pathological changes of synapses, first it is absolutely essential to know just what is ‘normal’. At the moment, we are still a long way from really grasping how the structure and function of synapses are connected,” states Lübke.

“At just a few micrometres, most synapses in the brain are so small that direct electrophysiological measurement is inconceivable for the foreseeable future. But we can create a quantifiable model and then specifically test certain synaptic parameters in normal and diseased brains in computer simulations,” the Jülich brain researcher explains.

It all started with the Calyx of Held

The idea of a model synapse in the brain can be traced back to Lübke’s collaboration with Bert Sakmann’s working group. In 1991, the physician shared the Nobel Prize with Erwin Neher for developing and using the patch clamp method, which allows the difference between electrical currents in biological cells to be measured very clearly. At the beginning of the new millennium, Lübke, Sakmann, and colleagues managed to create an initial three-dimensional, high-resolution, and quantifiable model of the Calyx of Held, a synapse named after the scientist who discovered it. This is a major junction found in the auditory pathway of many animal species, including humans. The synapse is by far the largest in the entire nervous system and has a diameter of 20 to 30 micrometres, making it around 3,000 times larger than the majority of synapses in the brain.

Held’sche Calyx3D reconstruction of the Calyx of Held (left, coloured yellow), which has over 600 active zones (neurotransmitter release sites), and the post-synaptic principle neuron (right, coloured blue)
Copyright: Journal of Neuroscience, Sätzler et al., DOI: (Copyright 2002 Society for Neuroscience)

The dream of deriving a generally applicable model synapse from the Calyx of Held did not last, however. It became apparent that the giant synapse is the exception rather than the rule and is adapted to the specific requirements of hearing. Subsequent studies emphasized how varied synaptic structures in different networks of the brain are in their constitution, with the mossy fibre bouton, a synapse involved in learning and memory processes, as an example. Therefore, not just one, but a whole range of synapse models is needed depending on which area of the brain researchers are interested in.

From animals to humans

There is also another problem. Practically all previous findings relating to the structure of synapses are based on results from animal models. Lübke and Sakmann also spent years working on reconstructing the Calyx of Held from the ultra-thin tissue sections of a rat. At that time, nobody could say how applicable the findings were to human synapses.

“Post-mortem brains, i.e. brains from deceased donors, which are commonly used for structure investigations, have a crucial drawback, and that is that an inordinate amount of time passes between when a donor dies and when a neuropathologist can remove their brain. During this time, the ultra-fine structures of the brain undergo a substantial transformation that leads to a dramatic loss of quality, at least at the cellular and sub-cellular levels. Therefore, post-mortem brains are unfortunately unsuitable for our studies of synapse structures," says Lübke.

He and his team achieved a breakthrough about eight years ago. They entered into cooperation agreements with the university hospital in Bonn and later the one in Bochum. The neurosurgeons there helped them to acquire fresh tissue samples for their research.

EM-PräparateUltra-thin sections of 50 nanometres in thickness are applied to copper racks of around half a millimetre in diameter for examination under an electron microscope.
Copyright: Forschungszentrum Jülich / Ralf-Uwe Limbach

The samples come from tumour patients who are operated on or people suffering from epilepsy. If chemotherapy or medication can no longer provide any help, surgery is the only option for removing diseased parts of the brain. The first step in epilepsy surgery is identifying the epileptic focus. If underlying brain structures such as the hippocampus are affected, the tissue above it in the cerebral cortex is first removed with careful precision in order to gain access. Lübke can – with the consent of the patient – use this tissue for his research once it is removed.

“When someone undergoes an operation, I am notified. I then go to Bochum with the complete set of chemicals that I need to prepare the biopsy material for examination with an electron microscope. That means that when the skull is opened and the access cerebral tissue is isolated and removed, I am practically standing right next to the surgeon and the tissue is then fixed immediately," says Lübke.

The cortical column at a glance

Over the next three and a half years, the Jülich brain researcher hopes to model the synaptic organization of a cortical column in the temporal lobe based on these samples, covering each of the six individual layers of the cerebral cortex. The temporal lobe is a multi-functional brain region that plays an important role in a variety of different senses. It is connected with hearing and sight, among other senses, but also with learning and memory processes.
A cortical column is the smallest functional module in all the sensory areas, typically stretching across every layer. After researching healthy tissues, Lübke is eager to study the structures in the pathologically altered human neocortex and hippocampus – these being areas where certain kinds of epilepsy are focused.

3D-Rekonstruktion Series of images from electron microscopes are composed virtually on a computer from over a hundred individual images in some cases to form a 3D model.
Copyright: Forschungszentrum Jülich

He and his team have already analysed healthy tissue from two of these cortical layers – layers 4 and 5 – and generated representative quantitative 3D models made up of hundreds of reconstructed synapses. The results were published in the renowned journals Cerebral Cortex and eLife, and in a review article in Neuroforum late last year.

Fundamental variations

“One surprising result was that synapses in the human cerebral cortex, along with having some similarities in certain structural parameters, differ substantially from those of their ‘relatives’ in animal models. In the regions of the brain that we have studied so far, the active zones that are so crucial for signal transduction are about twice the size in humans as they are in rats or mice, for example. Even the pools of synaptic vesicles – small bubbles filled with neurotransmitters – are five times bigger than those in the animal model in some cases," explains Lübke.

Vergleich von synaptischen Komplexen Comparison of synaptic complexes in human and murine (mouse) neocortices: (A) Single electron microscope frame through layer 4 of the neocortex in the temporal lobe of a human (yellow: synaptic bouton, blue: dendritic spine, red: active zone, green: synaptic vesicles). (B) 3D volume reconstruction of the synaptic complex shown in A. (C) Single electron microscope frame through a mouse’s neocortex. (D) Corresponding 3D volume reconstruction. In B and D, the size and shape of the active zone is further enlarged for the purpose of better visualization. Scale A–D 0.5 µm.
Copyright: Forschungszentrum Jülich

In addition to this, synapses from different layers also have clear fundamental differences from each other. This is shown by the results of the representative analysis conducted by the team  headed by Lübke. Layer 4 is commonly viewed as the cortical entry layer, which receives signals from the sensory periphery, for example from the optic or auditory nerve, and passes them on within the cortical column. Layer 5, by contrast, is seen as the most important exit layer, from which signals are forwarded to the sensory periphery over great distances.

The structures of synapses also turn out to be similarly varied, as the researchers have now shown. The synapses in layer 4 are much smaller than those in layer 5, but the active zones where the neurotransmitters are released are over twice the size. At the same time, synapses in layer 4 exhibit a significantly larger number of quickly available vesicles with neurotransmitters near the active zone, which cause them to behave differently in signal transduction.

“Obviously, the synapses in layer 4 have better transducation characteristics and therefore act as ‘amplifiers’ or even ‘discriminators’ – a kind of filter – for sensory signals, which they pass on to other layers within the cortical column. In contrast, the synapses in layer 5 collect and integrate signals across cortical columns and then send them back to the sensory periphery," says Lübke.

Difference between men and women

Not only are there differences between the different layers, there are also differences between the sexes. For a long time already there has been evidence that women’s brains are on average less voluminous, but have denser connections than those of men. Observations made by Lübke and his team regarding the density of synapses have provided further evidence for this theory, at least as far as the previously studied layers are concerned. They found out that the synapses in the cerebral cortex of women are packed together with two and half times the density of those in the cerebral cortex of men.

“These results suggest that women can achieve demonstrably better language skills, since the temporal lobe is also involved in processing memory and language, along with understanding language," says Lübke.

Tobias Schlößer


Joachim H. R. Lübke, Astrid Rollenhagen
Synapses: Multitasking Global Players in the Brain
Neuroforum (5 December 2019), doi:

Rachida Yakoubi, Astrid Rollenhagen, Marec von Lehe, Dorothea Miller, Bernd Walkenfort, Mike Hasenberg, Kurt Sätzler, Joachim HR Lübke
Ultrastructural heterogeneity of layer 4 excitatory synaptic boutons in the adult human temporal lobe neocortex
eLIFE (20 November 2019), doi:

Astrid Rollenhagen, Ora Ohana, Kurt Sätzler, Claus C. Hilgetag, Dietmar Kuhl, Joachim H. R. Lübke
Structural Properties of Synaptic Transmission and Temporal Dynamics at Excitatory Layer 5B Synapses in the Adult Rat Somatosensory Cortex
Front. Synaptic Neurosci. (30 July 2018), doi:

Rachida Yakoubi, Astrid Rollenhagen, Marec von Lehe, Yachao Shao, Kurt Sätzler, Joachim H R Lübke
Quantitative Three-Dimensional Reconstructions of Excitatory Synaptic Boutons in Layer 5 of the Adult Human Temporal Lobe Neocortex: A Fine-Scale Electron Microscopic Analysis
Cerebral Cortex (21 Juni 2018), doi:

Astrid Rollenhagen, Kerstin Klook, Kurt Sätzler, Guanxiao Qi, Max Anstotz, Dirk Feldmeyer, Joachim H.R. Lübke
Structural determinants underlying the high efficacy of synaptic transmission and plasticity at synaptic boutons in layer 4 of the adult rat ‘barrel cortex’
Brain Structure and Function (26 November 2015), doi:

Rollenhagen A1, Sätzler K, Rodríguez EP, Jonas P, Frotscher M, Lübke JH.
Structural determinants of transmission at large hippocampal mossy fiber synapses
Journal of Neuroscience (26 September 2007), doi:

Kurt Sätzler, Leander F. Söhl, Johann H. Bollmann, J. Gerard G. Borst, Michael Frotscher, Bert Sakmann, Joachim H. R. Lübke
Three-Dimensional Reconstruction of a Calyx of Held and Its Postsynaptic Principal Neuron in the Medial Nucleus of the Trapezoid Body
Journal of Neuroscience (15 December 2002), doi: