apl. Prof. Dr. Almut Schüz

Functional neuroanatomy of the cerebral cortex 

Almut Schüz

Brief summary of my research

Main topics (chronologically):

1) Mechanism of learning in the cerebral cortex

Deprivation experiments by some authors in the 60ies and 70ties suggested that learning leads to the formation of dendritic spines in the cerebral cortex. Also, a number of differences on the synaptic level were described after deprivation or enrichment. Most of these experiments where done on mice or rats, i.e.  altricial animals which are very immature at birth. I was therefore interested in the developmental state of the brain around birth in precocial animals which are born in a very mature state. Even in these animals, changes due to learning should happen mainly after birth. I investigated, therefore, the development of spines and synapses in the cortex of the guinea pig. It turned out that already before birth the density of dendritic spines and synapses reaches the adult values. Conclusion: the formation of spines and synapses in the cerebral cortex is not the result of learning but a prerequisite for it. An already richly connected network is necessary for the possibility of associative learning. (Schüz, 1978; 1981a).

This result suggests that the mechanism of learning is mainly due to modifications of already existing synapses and/or dendritic spines. Comparison between prenatal and adult guinea pigs showed differences in the average diameter of dendritic spines, as well as in the average number of presynaptic vesicles, making these features candidates for learning mechanisms. (Schüz, 1980, 1981b, 1986, 1988)

It remained open to what degree a turnover of the existing spines and synapses plays a role in learning. What we could observe that – in spite of a similar density of dendritic spines in prenatal and adult guinea pigs - there is still a slight postnatal increase and then decrease of dendritic spines, indicating an elimination of useless connections.

2) Network structure of the cerebral cortex

In the 80ies I joined Valentino Braitenberg in his quantitative investigations on the cerebral cortex. Together also with Günther Palm and other co-workers (see below), we quantified in the mouse cortex the density of neurons, types of neurons, dendritic spines, synapses, types of synapses, length density of axons and dendrites, etc. From these numbers, others could be derived like total numbers, numbers of spines and synapses per neuron and probabilities of one or more connections between neurons, depending on their distance. The results showed that the cortex is – in contrast to other parts of the brain – connected mainly excitatorily, that individual neurons are connected to thousands of others, mainly via spine synapses, and make both, local and distant connections, thus connecting each neuron to each other over only few synaptic steps. Thus, the cerebral cortex can be best interpreted as an associative network, and our results strongly support the Hebbian theory of learning. The individual results can be found a.o. in Schüz and Münster, 1985; Krone, Mallot, Palm, Schüz, 1986; Schüz and Dortenmann, 1987; Schüz and Palm, 1989; Schüz 1992; Braitenberg and Schüz 1992; Hellwig, Schüz, Aertsen; 1994; Schüz 1994; Aertsen, Erb, Palm Schüz (1994); Schüz 1995.

For a summary of this research see the book by Braitenberg and Schüz (1991, 1998).

The topic of brain size and connectivity is also treated in Braitenberg and Schüz (1998), as well as in Schüz and Sultan (2009).

For a comparison between cortical areas see the book edited by Schüz and Miller (2002).

3) Organization of cortico-cortical connections

Long connections via the white matter connect cortical areas over long distances in the human brain. However, quantitatively connections to neighouring and closely located areas prevail (Schüz and Braitenberg, 2002). The quantitative dominance of neighourhood connections was also found in a systematic tracer study on the mouse cortex (Schüz, Chaimow, Liewald, Dortenmann, 2005). For the total length of distant axonal arbours see Turesson und Schüz (2006).

In large brains, three types of connections exist in many pyramidal cells: 1) short-range connections in the vicinity of the neuron, 2) middle-range connections over distances of up to a few millimeters via long axon collaterals and 3) long-range connections via the white matter (Voges, Schüz, Aertsen, Rotter, 2010).

 4) Axon diameters (i.e. conduction times) in long-range cortico-cortical connections

Our studies in mice, monkeys and humans showed a similar distribution of axon diameters in  the cortical white matter of these species. Most axons in monkeys and humans were as small in caliber as those in the mouse (<0.001 mm inner diameter), i.e. must have slow conduction velocities. Only a small fraction of axons in monkey and humans are of large and very large caliber. Thus, there must be a large range of conduction times over the cortex in large brains, with much longer conduction times in average than in the mouse. This finding supports the idea of the existence of sequence detectors (Miller, 1996) in large brains as a possible basis for new kinds of functions, like for example language. (Liewald, Miller, Logothetis, Wagner, Schüz; 2014).

Further topics

Hippocampus: Commonalities and differences between the structure of neocortex and hippocampus (Braitenberg and Schüz, 1983).

Cerebellum: Comparison of its development in a precocial and altricial rodent (Schüz and Hein, 1984)

Septal nuclei: Investigation on human brains from New Guinea (Koch, Schüz and Kariks, 1985), motivated by a study by Beck and Gajdusek (1996) on Papua brains in connection with Kuru disease.

Development of dendritic spines and synapses in organotypic cell cultures: The development in such cultures is very similar to the development in situ (Caeser and Schüz, 1992).

What is first: the dendritic spine or its synapse? Our results suggest that the synapse is developed first and then the spine grows out (Schweizer and Schüz, 1990)

Comparison of spines and synapses in different species (hedgehog and monkey): There was no difference in density and size of synapses and proportion of excitatory and inhibitory synapses. Average spine density was slightly higher in the macaque. Spine heads in monkey were often largely surrounded by their presynaptic element. (Schüz and Demianenko, 1995)

Visual system of primates: investigations on V1 and V2 of the new world monkey Callithrix; doctoral thesis of Matthias Valverde Salzmann. He could considerably improve the method for localization of functions (optical imaging) onto the postmortem anatomy of the brain. This made it possible to localize colour vision in these animals. In contrast to old world monkeys, colour vision exists nearly only in female animals. It could be shown that in spite of this evolutionary difference, colour vision in female Callithrix is localized similarly as in old world monkeys. (Valverde Salzmann, Wallace. Logothetis, Schüz, 2011; Valverde Salzmann, Bartels, Logothetis, Schüz, 2012). 

Cooperations

Plasticity of the cortex after retinal lesion: It turned out that plasticity of V1 after a retinal lesion is limited. MRI experiments in the group of Nikos Logothetis showed that the projection zone in V1 of the lesioned spot in the retina did not respond to visual stimuli anymore even after 7 ½  months. (Smirnakis, Brewer, Schmid, Tolias, Schüz, Augath, Inhoffen, Wandell, Logothetis, 2005).

Visualisation of the Tracer Biocytin in MRI: The chemists in Logothetis’ group combined the tracer biocytin with Gadolinium, so that it became visible in MRI and they further modified the molecule such that it became stable over longer time spans than the biocytin used in histology. By way of histological methods the degree of uptake and transport was investigated and showed that this approach can indeed be used to visualize projections in vivo. (Mishra, Schüz, Engelmann, Beyerlein, Logothetis, Canals, 2011; Mishra, Mishra, Canals, Logothetis, Beyerlein, Engelmann, Schüz, Dhingra, 2012).   

Development of a molecular learning network: In this EU-Project Marco Fontana and his group at the University of Parma used conducting molecules in order to develop a network the conductivity of which changes, depending on its use, thus imitating the plasticity of the nervous system. The aim is to develop a new kind of hardware which is fundamentally more brainlike than conventional computers. (Erokhin, Schüz, Fontana, 2009; Erokhin, Berzina, Gorshov, Camorani, Pucci, Ricci, Ruggeri, Sigala, Schüz, 2012; Sigala, Smerieri, Camorani, Erokhin, 2013).

Vascularisation of cytochromeoxidase blobs in the visual cortex of primates: cytochrome-oxidase blobs had been reported to be considerably more vascularized than interblob regions. However, they could not be visualized in high-resolution MRI. Because of these contradictory findings we re-investigated the density of vascularization in blob and interblob regions with a different histological method than in the previous study. We could confirm a higher density of vascularization, however, the difference was one order of magnitude less than reported in the previous study. (Keller, Schüz, Logothetis, Weber, 2011).