For an extended, technical description of this field, see this review.
For some friendly press coverage, check out the Mason Gazette report.
Otherwise... enjoy this page!

[1] L-Neuron: Generation and Description of Dendritic Morphology

The primary goal of the L-Neuron project is to create virtual neurons that are anatomically indistinguishable from their real counterparts. L-Neuron uses the formalism of the Lyndenmayer systems in the version known as Turtle Graphics. Particularly, the L-Neuron program is based on a modification of Laurens Lapre's "L-Parser". Instead of working with the standard L-systems algorithms, however, L-Neuron implements sets of neuroanatomical rules discovered by several research groups (and in particular, Hillman's, Tamori', and Burke's). These rules are local and recursive.

The fact that the rules are local means that they establish correlations among geometrical parameters (e.g. a dendritic branch's diameter and taper) independent of their overall position in the tree. The term recursive refers to the fact that, as a branch grows, it stems other branches that follow the same rules. Therefore the same simple algorithm can be reiterated many times as the dendritic tree develops more and more bifurcations (as shown in this figure, representing the Hillman rules). The L-Neuron algorithms read in experimental data to generate virtual structures. The experimental data are in the form of statistical distributions (for example, bifurcation angles in Purkinje cells can be represented with a Gaussian distribution, with a certain average and standard deviation). L-Neuron samples the values of the parameters within these statistical distributions in a stochastic (random) fashion during dendritic growth. Therefore, with the same set of parameter distributions, the program can generate an unlimited number of virtual neurons.

Because L-Neuron implements a stochastic and statistical algorithm, it achieves a form of morphological data compression and amplification. Compression because hundreds of experimental neuronal tracings within a certain morphological class can be described with a handful of statistical distributions, instead of with the classical compartmental description (thousands of lines per neuron). Amplification because, from those hundred real neurons, one can then generate thousands of virtual counterparts.

Note that different sets of statistical distributions correspond to different morphological families. By varying these values, the same algorithm can describe neurons as diverse as pyramidal, granule, Purkinje, or stellate cells. In this screen shot of the L-Neuron viewer, a pyramidal cell was generated with experimental values reported by Hillman. Green segments are basal dendrites, blue segments are apical dendrites, and the soma is drawn in red.

L-Neuron can output its virtual structures in a variety of formats, including virtual reality, graphical files, and the standard neuroanatomical coordinates compatible with neurophysiological simulators such as GENESIS and Neuron.
 

If you are interested in this project, please visit the L-Neuron Web page...

The L-Neuron Project is supported by the Krasnow Institute for Advanced Study and by Human Brain Project grant R01-NS39600-01 awarded to Giorgio Ascoli by the National Institute of Neurological Disorders and Stroke (NIH).
 

[2] Influence of Dendritic Morphology on Neuronal Electrophysiology

Why is it so important to model neuronal anatomy in detail? Neuroscientists are convinced that dendritic morphology plays an important role in neural computation, but there are very few attempts to investigate this role quantitatively. We are systematically studying the effect of the geometry and topology of neurons on their electrical behavior by means of computational simulations.
We took several experimentally traced neurons from a public electronic archive, converted them for use with the GENESIS simulator, and loaded them with a standard model for their morphological class (CA3 pyramidal cell's Traub model). We paid special attention in setting the exact same distributions of electrophysiological properties (e.g. ionic concentrations and conductances) in all the neurons. Then we started stimulating them (with somatic current injections) with an identical protocol for all of the cells. Thus, every single parameter was constant across these neurons, aside from their dendritic morphology. This variability was sufficient to cause both qualitative and quantitative differences in the firing output of the neurons. Different qualitative behaviors were observed as distinct types of firing modes (regular spiking, as for the two top cells, or train bursting, as in the two bottom cells). Within each mode, there were dramatic quantitative differences (as in the spiking rate between the two top cells, or in the baseline within a burst, in the bottom cells). This project aims at quantifying the exact relationships between morphological and physiological parameters.

Recently, we started applying this research method to the study of the involvement of dendritic morphology in the abnormal neuronal electrophysiology at the basis of Alzheimer's Disease, in a project supported by the Krasnow Institute and by Award No. 00-1 from the Commonwealth of Virginia’s Alzheimer’s and Related Diseases Research Award Fund, administered by the Virginia Center on Aging, Virginia Commonwealth University.
 

[3] Anatomically Accurate Neural Networks: Building a Hippocampus

Can virtual neurons be assembled in realistic neural networks, and can these be used to study the electrophysiological behavior at the system level? Steve Senft has developed a program, called ArborVitae (AV), that implements stochastic and statistical algorithms similar to those described for L-Neuron at a population level.

As an example of the ArborVitae output, here we show the main cells of the rat hippocampus. In each panel, the upper four neurons are real cells from the Southampton archive. The lower four neurons are created with AV. Axons are not present in any of the cells. Each of the AV cell has only approximately 1/10 of the dendritic compartments of a corresponding real neuron. Upper left panel: CA3 pyramidal cells. Color code: basal dendrites are brown (receiving inputs from gabaergic interneurons, cholinergic septohippocampal pathway and glutamatergic Schaffer collaterals), proximal apical dendrites are green (receiving inputs from gabaergic interneurons, glutamatergic mossy fibers and Schaffer collaterals), distal apical dendrites are blue (receiving inputs from gabaergic interneurons, glutamatergic perforant pathway and Schaffer collaterals). In the real AV model the distal apical dendrites are more sharply oriented towards the top (away from the basal dendrites). Here this effect is diluted by the lower density of basal dendrites (only four neurons are present!). Upper right panel: CA1 pyramidal cells. Color code: basal dendrites are brown (receiving inputs from gabaergic interneurons, cholinergic septohippocampal pathway and glutamatergic CA1 axonal collaterals), apical dendrites are green (receiving inputs from gabaergic interneurons and glutamatergic Schaffer collaterals). Lower left panel: DG granule cells. The dendrites receive their inputs from gabaergic and glutamatergic interneurons, cholinergic septohippocampal pathway and glutamatergic perforant pathway. Lower right panel: polymorphic cells. This stellate-like structure is adopted by several neuronal families such as GPC and mossy cells in DG, Oriens interneurons in CA3 and Alveus interneurons in CA1.

ArborVitae also implements an algorithm to describe axonal navigation and synaptic connectivity. We took advantage of this feature to generate a virtual, small-scale model of a hippocampal slice. This structure consists of the dentate gyrus granule cell layer (bottom right in the figure), the CA3 and CA1 pyramidal cell layers (left and top right in the figure, respectively), as well as an off-field "black-box" entorhinal cortical module sending axons to the granule cells and receiving axons from CA1, and a septohippocampal input to CA3. Because this network is interconnected, we were able to simulate a simple form of electrical transmission (white colors indicate depolarized membranes). We are now working on a larger-scale model of the hippocampal slice. If you want to learn more, please see our technical report "Computational Neuroanatomy of the Hippocampus".

Our interest in the hippocampus is motivated by several reasons: [A] The hippocampus is involved in associative learning, one of the basic building blocks of mammalian higher cognitive functions. [B] The rat hippocampus is among the best known neuroanatomical structures, and morphological data are extensively available in the scientific literature. [C] The hippocampus has a mainly lamellar structure, therefore an entire hippocampus can be assembled by stacking together many slices. In other words, the system is easily scalable up in the computational model. We ran an extensive literature search of the cellular connectivity of the rat hippocampus, and this is the basis of our larger scale anatomical model.
 

[4] QRN: A New Formalism for the Efficient Simulation of Neuronal Electrophysiology

A realistic anatomical model of an entire portion of a mammalian nervous system contains thousands of neurons, and millions of compartments. It is currently impossible to simulate the electrophysiological activity in such a large-scale system with standard computers and the classic formalism of differential equations (such as implemented in the GENESIS simulator). In order to overcome this problem, we have investigated the possibility to use qualitative reasoning as a computationally efficient approach to neurophysiological modeling. The Qualitative Reasoning Neuron (QRN) is a tool developed by Jeff Krichmar to compute and describe the electrophysiological activity of neurons, down to the molecular level, without the need to solve any differential equation. QRN also implemented additional "shortcuts" to optimize dendritic simulations, such as a non-Cartesian coordinate system for dendritic spines, and an event-driven timing system. This formalism allowed us to run simulations with up to half a million of compartments (such as in our model of the cerebellar cortex) on regular desktops. If you want to learn more about QRN, visit the QRN web page.
 

Back to the home page of the Computational Neuroanatomy Group
 

Publications

[Mason Gazette Report covering our research.]

1) G. Ascoli, R.F. Goldin: Coordinate systems for dendritic spines: a somatocentric approach. Complexity 2(4):40-48 (1997).

2) J.L. Krichmar, G. Ascoli, L. Hunter, J.L. Olds: A model of cerebellar saccadic motor learning using qualitative reasoning. Lecture Notes Computer Science 1240:134-145 (1997).

3) J.L. Krichmar, G. Ascoli,  J.L. Olds, L. Hunter: Qualitative reasoning as a modeling tool for computational neuroscience. In: Computational Neuroscience: Trends in Research 1998, J.M. Bower Ed., 609-614, Plenum NY. (1998).

4) J.P. Vandersluis, J.D. Cooke, G. Ascoli, J.L. Krichmar, G. Michaels, M. Montgomery, J. Symanzyk, B. Vitucci: Visualization of high-dimensional biomedical data. Comp. Sci. Stat. 30:482-487 (1998).

5) S.L. Senft, G. Ascoli: Reconstruction of brain networks by algorithmic amplification of morphometry data. Lect. Notes Comp. Sci. 1606:25-33 (1999).

6) J. Symanzik, G. Ascoli, S. Washington, J. Krichmar: Visual data mining of brain cells. Comp. Sci. Stat. 31:445-449 (1999).

7) G. Ascoli: Progress and Perspectives in computational neuroanatomy. Anatom. Rec. 257(6):195-207 (1999).

8) G. Ascoli, J.L. Krichmar: L-Neuron: a modeling tool for the efficient generation and parsimonious description of dendritic morphology. Neurocomputing 32-33:1003-1011 (2000).

9) S. Washington, G. Ascoli, J. Krichmar: Statistical analysis of CA3 pyramidal cell morphology’s effect on electrophysiology. Neurocomputing 32-33:261-269 (2000).

10) S. Nasuto, J. Krichmar, R. Scorcioni, G. Ascoli: Algorithmic analysis of electrophysiological data for the investigation of structure-activity relationship in single neurons. InterJournal Complex Syst. 389:1-7 (2000).

11) R. Scorcioni, G. Ascoli: Algorithmic extraction of morphological statistics from electronic archives of neuroanatomy. Lect. Notes Comp. Sci. 2084:30-37 (2001).

12) S. Nasuto, J. Krichmar, G. Ascoli: A computational study of the relationship between neuronal morphology and electrophysiology in an Alzheimer's Disease model. Neurocomputing 38-40:1477-1487 (2001).

13) G. Ascoli G., J. Krichmar, S. Nasuto, S. Senft: Generation, description and storage of dendritic morphology data. Phil. Trans. R. Soc. B, 356(1412):1131-45 (2001).

14) G. Ascoli, J. Krichmar, R. Scorcioni, S. Nasuto, S. Senft: Computer generation and quantitative morphometric analysis of virtual neurons. Anat. Embryol. 204:283-301 (2001).

15) G. Ascoli,  A. Samsonovich: Bayesian morphometry of hippocampal cells suggests same-cell somatodendritic repulsion. Adv. Neural Inf. Proc. Syst. 14:133-139 (2002).

16) G. Ascoli (Ed.): Computational Neuroanatomy: Principles and Methods (Humana Press, Totowa, NJ, 2002).
        Chap. 1 (pp 3-26): Computing the brain and the computing brain (G. Ascoli)
        Chap. 3 (pp 49-70): Generation and description of neuronal morphology using L-Neuron: a case study. (D. Donohue, R. Scorcioni, G. Ascoli)
        Chap. 6 (pp 105-125): The relationship between neuronal shape and neuronal activity. (J. Krichmar, S. Nasuto)
        Chap 7 (pp 127-148): Practical aspects in anatomically accurate simulations of neuronal electrophysiology. (M. Lazarewicz, S. Boer-Iwema, G. Ascoli)
        Chap 12 (pp 245-270): Axonal navigation through voxel substrates: a strategy for reconstructing brain circuitry. (S. Senft)
        Chap 19 (pp 425-436): Towards virtual brains. (A. Samsonovich, G. Ascoli)

17) R. Scorcioni, J.-M. Boutiller, G. Ascoli: A real scale model of the dentate gyrus based on single-cell reconstructions and 3D rendering of a brain atlas. Neurocomputing, 44-46:629-634 (2002).

18) G. Ascoli: Neuroanatomical algorithms for dendritic modeling. Network: Comput. Neural Syst. 13:247-260 (2002).

19) J. Krichmar, S. Nasuto, R. Scorcioni, S. Washington, G. Ascoli: Influence of dendritic morphology on CA3 pyramidal cell electrophysiology. Brain Res., 941:11-28 (2002).

20) Turner DA,  Cannon RC, Ascoli GA: Web-based neuronal archives: neuronal morphometric and electrotonic analysis. In R. Kotter (Ed.): Neuroscience Databases – A Practical Guide, 81-98, Elsevier, Amsterdam (2002).

21) M. Lazarewicz, M. Migliore, G. Ascoli: A new bursting model of CA3 pyramidal cell physiology suggests multiple locations for spike initiation. Biosystems, 67:129-37 (2002).

22) Ascoli G.: Electrotonic effects on spike response model dynamics. IEEE Neural Networks, 2831-2836 (2003).

23) Samsonovich A., Ascoli G.: Statistical morphological analysis of hippocampal principal neurons indicates selective repulsion of dendrites from their own cell. J. Neurosci. Res. 71:173-87 (2003).

24) Ascoli G.: Passive dendritic integration heavily affects spiking dynamics of recurrent networks. Neural Networks, 16:657-663 (2003).

25) Scorcioni R., Lazarewicz M., Ascoli G.: Quantitative morphometry of hippocampal pyramidal cells: differences between anatomical classes and reconstructing laboratories. J. Comp. Neurol., 473:177-193 (2004).

26) Ascoli G., Atkeson J.: Incorporating  anatomically  realistic  cellular-level  connectivity  in  neural  network  models  of  the  rat  hippocampus. Biosystems, 79:173-181 (2005).

27) Li Y., Brewer D., Burke R.E., Ascoli G.: Developmental changes in spinal motoneuron dendrites in neonatal mice. J. Comp. Neurol., 483:304-317 (2005).

28) Donohue D., Ascoli G.: Models of neuronal outgrowth. In Koslow S.H. and Subramaniam S. (Eds.): Databasing the Brain: From Data to Knowledge, Wiley, New York, NY. pp 303-323 (2005).

29) Samsonovich A., Ascoli G.: Statistical determinants of dendritic morphology in hippocampal pyramidal neurons: a hidden Markov model. Hippocampus, 15:166-183  (2005).

30) Samsonovich A., Ascoli G.: A simple neural network model of the hippocampus suggesting its pathfinding role in episodic memory retrieval. Learning & Memory, 12:193-208  (2005).

32) Samsonovich A., Ascoli G.: Algorithmic description of hippocampal granule cell dendritic morphology. Neurocomputing, 65-66:253-260 (2005).

33) Donohue D., Ascoli G.: Local diameter fully constrains dendritic size in basal but not apical trees of CA1 pyramidal neurons. J. Comput. Neurosci., 19:223-238 (2005).

34) Brown K., Donohue D., D’Alessandro G., Ascoli G.: A cross-platform freeware tool for digital reconstruction of neuronal arborizations from image stacks. Neuroinformatics, 3:343-360 (2005).

35) Migliore M., Ferrante M., Ascoli G.: Signal propagation in oblique dendrites of CA1 pyramidal cells. J. Neurophys., 94:4145-4155 (2005).

36)  Samsonovich A., Ascoli G.: Morphological homeostasis in cortical dendrites. Proc. Natl. Acad. Sci. USA, 103:1569-1574 (2006).

37) Ascoli G: Mobilizing the base of neuroscience data: the case of neuronal morphologies. Nature Rev. Neurosci., 7:318-324 (2006).

38) Ascoli G., Scorcioni R.: Neuron and network modeling. In Zaborszky L., Wouterlood, Lanciego J. (Eds.): Tract Tracing Methods, 3rd ed., Springer, New York NY. pp. 604-630 (2006).

39) Li X., Ascoli G.: Computational simulation of the input-output relationship in hippocampal pyramidal cells. J. Comput. Neurosci., 21:191-209 (2006).

40) Krichmar J., Velasquez D., Ascoli G.: Effects of beta-catenin on dendritic morphology and simulated firing patterns in cultured hippocampal neurons. Biol. Bull., 211:31-43 (2006).

41) Samsonovich A., Ascoli G.: Computational models of dendritic morphology: from parsimonious description to biological insight. In Costa L.da F. and Mueller G. (Eds.): Modeling Biology: Structures, Behaviors, Evolution, MIT Press, Cambridge, MA. In Press (2006).

Back to the home page of the Computational Neuroanatomy Group