Explaining the title

10. Instead of a conclusion: explaining the title.

To me personally this interview is a physical evidence of failure of scientific program, which is clearly suffering from substitution and “displacement” of science by technology.

I already discussed how complexity of modern technology can be abused by misleading terminology. The second opportunity complex technology creates is the easy substitution of scientific results by purely technological. It is well known move to conclude technology report with promise of future potential for some research: in the mind of many it magically transforms the paper to scientific. Unfortunately, this appearance does not last too long.

A model or not a model, but BigBrain data set is useful for two purposes. First it is a potential source of microscopy data for cytoarchitectonic studies in 3D. Second – it is a template for visualization of other brain-related data. As a template it can be used already, but that is not the case with the first purpose.

Regretfully, after almost 17 years, 1-micron resolution data are still not published in BigBrain data set, and scientific results of research of BigBrain cytoarchitectonic are very scarce, to say the least. (Two results regarding cortical thickness gradient (Wagstyl K. et al.) and similarity of profiles for connected areas (Wei Y et al.) are base on 20 mk resolution and will be discussed in my next post). Of course, if the technological part of the work, including the development of programs and algorithms, preparation of sections, manual correction of the images, etc. takes so much time, then there is no other way out but to report purely methodological results. But, perhaps, such a long protracted technological phase is explained by something other than the complexity of the technology? What if the reason for such situation is an irrationally-defined goal?

In particular, no matter how attractive is the idea of showing the entire human brain inside a computer exactly as it is visible in the microscope, maybe one shouldn’t try to do this for the whole brain all at once? After all, we know from the history of cytoarchitectonics that people did not immediately learn how to make slices of the whole brain. First, small blocks were serially cut. Perhaps, it would be possible to achieve registration at a resolution of 1 micron of relatively small blocks, put them on a disk and make these data blocks publicly available for research.

This way, it would be possible to accumulate the material necessary for cytoarchitectonic studies of different fields of the cortex in 3D in a reasonable time and to obtain scientifically significant results in parallel with an increase in the volume of data blocks. An alternative to this situation would be waiting for data with a resolution of 1 micron for another 15 years, which would result in even more speculative presentations and reporting. Evidently, the patience of funding agencies would burst rather sooner than later.

Moreover, how is it even feasible to create a map with 1 micrometer resolution, if mapping technique available so far has an error of at least 200 micron theoretically, and much higher in reality, as we could see for ourselves? What sense does it make to ignore the biological “uncertainty principle” of position of area borders, clearly demonstrated in all areas mapping papers of K. Zilles, K. Amunts et all.[see references to this blog], and to pretend that the map with 1-micrometer resolution has such an important role in understanding “the whole brain”? Additionally, one should realize, that we did not even start to consider the problem of individual variability, which makes the whole story about 1-micron resolution map look like speculation, and nothing more.

So, we clearly collected enough evidence to demonstrate that BigBrain project is a pure magic: it managed to make visible single neurons in 20 micrometers-resolution images, which is physically impossible. We demonstrated that fuzzy shadows are visible in those images instead of real neurons. We also demonstrated how project’s participants are unanimously pushing the importance of the access to 1 micrometer-resolution data, exactly for this very purpose: to look at the whole brain at the level of single neuron. But unfortunately, such so clearly visible goal of the project, declared years ago, is still inaccessible and invisible even on the horizon. Of course, whatever is to blame, including the volume of data, slow download speed, the difficulties to support several users at once, and much more. In a word – everything except obvious reluctance of the project management to share uninterrupted series of brain section files, perhaps – even unregistered, where neurons will be visible through the whole brain without additional exaggerations and reservations.

Because, as we see it, it would make sense to project some functional MRI image on 1 micrometer-resolution “model” to see what is the architecture at neuronal level of the underlying area of cortex. However, if such model does not show neuronal organization, the whole idea becomes fruitless. Practically speaking, it makes very little difference what template to use for such a projection: anatomical template or MRI-based template (Collin27, MNI152 or alike). If there are no neurons visible – the whole game falls apart.

As for using BigBrain data set  as a template for visualization, I have a suggestion. Practically speaking, 3D templates up to 100 micrometer resolution available today are more than enough for this purpose. But if better resolution templates are really necessary, while high-resolution data is “on its way”, let’s temporarily use matryoshkas of different sizes as a templates for visualizing brain maps (Fig.9). The higher the data resolution, the bigger matryoshka should be used. Thankfully, nested dolls have always been of different sizes, and they are very easy to store: they fit perfectly into one another and do not take up much space. They also might be used as a symbolic gesture  to pay respect to Russian connections, hidden ever so remotely in deep historic roots of cytoarchitectonic research. Let’s call them HIBALL matryoshkas!

Fig. 9. Using matryoshkas templates to visualize cortical maps of increasingly high resolution.

  1. Katrin Amunts et al. “BigBrain: An Ultrahigh-Resolution 3D Human Brain Model”, Science 340, 1472 (2013); DOI: 10.1126/science.1235381.
  2. Liron Pantanowitz, et al. “Twenty Years of Digital Pathology: An Overview of the Road Travelled, What is on the Horizon, and the Emergence of Vendor-Neutral Archives”. J Pathol Inform. 2018; 9: 40.
  3. Hartmut Mohlberg, Bastian Tweddell, Thomas Lippert, and Katrin Amunts. “Workflows for Ultra-High Resolution 3D Models of the Human Brain on Massively Parallel Supercomputers”.  K. Amunts et al. (Eds.): BrainComp 2015, LNCS 10087, pp. 15–27, 2016.
  4. A. Schleicher, K. Amunts, S. Geyer, P. Morosan, K. Zilles. “Observer-independent method for microstructural parcellation of cerebral cortex: a quantitative approach to cytoarchitectonics”, NeuroImage, 9, 165-177, (1999).
  5. Istomin, V. “Automatic morphocorticography: development of the method and application results”. Dissertation of the Candidate of Medical Sciences, Moscow, 1987 (Rus.)
  6. Istomin V. Quantitative cytoarchitectural analysis of some cerebral cortical areas using the method of automated morphocorticography (Eng.). Neuroscience and Behavioral Physiology 1988;18(4):344-349.
  7. Блинков С.М., Глезер И.И. Мозг человека в цифрах и таблицах. Л.: Медицина, 1964. — 472 с. (Blinkov S.M., Gleser, I.I. Human brain in figures and tables. Leningrad, Medicina, 1964, 472 p.).
  8. V S Kesarev. “Quantitative architectonics of the human brain, Vestn Akad Med Nauk SSSR, 1978;(12):29-36.
  9. Кесарев В. С. “Пространственная организация новой коры мозга китообразных и приматов в сравнительном онтогенезе” : Дис… д-ра мед. наук : 03. 00. 11 / Виллен Семенович Кесарев. — М., 1986 г. — 270 с. : ил., табл.. (Kesarev V.S. “Spatial organization of the neocortex of cetaceans and primates in comparative ontogenesis”. Dissertation of the Doctor of Medical Sciences. Moscow, 1986, 270 pp.)
  10. Деревягин В. И., Левченкова В. Д., Кесарев В. С., Борисенко О. В. “Количественный анализ структурных изменений прецентральной области коры мозга человека при детском церебральном параличе”. Сборник научных трудов : 1990. — Вып. 19 : Макро- и микроуровни организации мозга. — С. 132-134. (Quantitative analysis of structural changes in human precentral cerebral cortex in infantile cerebral palsy).
  11. Поляков Г.И. “Основы систематики нейронов новой коры большого мозга человека”, “Медицина”, Москва, 1973. (Poliakov, G.I. “Fundamentals of systematic of neurons of neocortex of the cerebrum of human brain”. “Meditsina”,  Moscow, 1973.)
  12. Леонтович Т.А. “Нейронная организация подкорковых образований переднего мозга”. Москва, Медицина, 1978, 384 стр. (Leontovich T.A. “Neuronal structure of Subcortical Formations of the Forebrain”, M., “Meditsina”, 1978, 384 pp., ill.)
  13. “Improving Cytoarchitectonic Segmentation of Human Brain Areas with Self-supervised Siamese Networks” H.Spitzer, K.Kiwitz,K.Amunts, S.Harmeling, T. Dickscheid. Proceedings of 21st International Conference MICCAID, Granada, Spain, 2018.