## Abstract

The body of knowledge about the connectivity of brain networks on different structural scales is growing rapidly. This information is considered highly valuable for determining the neural organization underlying brain function, yet connectivity data are too extensive and too complex to be understood intuitively. Computational analysis is required to evaluate them. Here we review mathematical, statistical, and computational methods that have been used by ourselves and other investigators to...

## Info

Column labeled Real CA1 refers to the 24 traced hippocampal cells obtained from the Duke-Southampton archive. Column labeled Unedited refers to cells created using the raw extracted data. The Edited column refers to cells created when negative tapers were not included in the data analysis, and maximum PK value was set at 2. One of these 50 cells was several orders of magnitude larger than the others, and the No Outlier column summarizes the data on the remainder cells. See Glossary in the...

## F fc

Ocular dominance maps produced by the EN in two dimensions. Cortical points are colored white or black depending on the eye to which they are committed. The two retinal sheets were hexagonal grids with a circular boundary, and the cortical sheet has a hexagonal grid with an elliptical boundary. (A) Normal development. (B) Strabismic development (reduced correlation between the eyes). (C and D) Effects of monocular deprivation. (C) Activity in the deprived eye reduced by 25 . (D)...

## Elb

Babbage, Charles, 426 Bardeen, John, 426 Basal scatter plot, 117f initial tree diameter, 58f parameters tree type parameters, 63t Basal forebrain, 12 afferents differential distribution, 184f CGRP-containing axons, 185f cholinergic dendritic segments, 185f cholinergic neurons, 173f, 176f compartments, 172 composition, 171 corticopetal system afferent restricted localization, 183 animals, 188 171-194 connections probability, 183-186 data acquisitions, 188-189 data analysis, 190-194 hodologically...

## Preface

The importance of computational modeling as a research approach in neuroscience is recognized today by most researchers in the field. Computational neuroscience is generally associated with simulations in electrophysiology and neural dynamics. Recently, an increasing number of neuroscientists have begun to use computer models to study and describe neuroanatomy, its subcellular bases, and its relationship with neuronal activity and function. Other researchers began importing accurate and...

## Association And Segregation Of Different Hodologically Identified Neural Populations

Although there is considerable species variation in the precise locations of cholinergic projection neurons in the BF, the efferent projections of these cells follow basic organizational principles in all vertebrate species studied. Thus in rodents, neurons within the medial septum and nucleus of the vertical limb of the diagonal band provide the major cholinergic innervation of the hippocampus cholinergic cells within the horizontal limb of the diagonal band project to the olfactory bulb,...

## Using Neurosys To Study Emergent Properties Of Neuronal Ensembles

The general approach we have used to study emergent properties of the cercal sensory system is to query the database for a set of neurons and their attributes and then to use the computational and visualization tools in NeuroSys to test predictions about system function. In all cases, these predictions cannot be made by studying individual neurons in isolation, but emerge when the attributes of many neurons are studied in combination. In the following sections, this process will be illustrated,...

## Trygve B Leergaard md phd

Trygve Leergaard is a postdoctoral research fellow at the Neural Systems and Graphics Computing Laboratory at the Institute of Basic Medical Sciences at the University of Oslo. He is experienced with computerized, three-dimensional anatomical analyses of developing and adult cerebro-cerebellar systems in rat, with a main focus on somatosensory map transformations.

## Cholinergic Cell Groups Show Regionally Selective Dendritic Orientation

Mean 3D Vector of Dendritic Processes Since the geometry of axons and dendrites imposes constraints on their connections, in order to understand how information is handled in the BF, it is important to determine how the shape of the axonal and dendritic arborizations could influence regional connectivity patterns. Cholinergic cell bodies give rise to 2-5 primary dendrites radiating in all directions. The relatively straight primary dendrites bifurcate in an iterative fashion, and the sum...

## J

D Search In the currently connected Workspace Database P Search In my personal Workspace Database Found one match- Double-click on the result to inspect it D Search In the currently connected Workspace Database P Search In my personal Workspace Database Found one match- Double-click on the result to inspect it

## Conclusion

Structurally realistic neuronal models can serve as devices to collect, evaluate, and distribute information concerning the functional organization of nervous systems. As we have described in this chapter, the central goal of the Modeler's Workspace project is to provide the neuroscience community with a modular, extensible, and open software environment enabling neuroscientists to develop, use, and share structurally realistic models. The purpose of this chapter is to describe a design that we...

## S

-jr2 G (t) dt'' D G (t) dtfdt' Eq. 3 where G is a gradient vector (orientation of diffusion measurement). To solve this equation, we need S0 that corresponds to the image intensity without diffusion weighting (the left most image in the upper row of Fig. 7) and at least six images (S) with different G (diffusion weighted images in Fig. 7). This can be more easily understood from the visual presentation in Figure 8. The radius of a diffusion ellipsoid along a particular axis represents the...

## F

Factor analysis, 313-316 Fast spin echo (FSE) MRM, 391 FDP algorithm, 78-80 Ferret neural activity perturbed patterns, 353 Conel data, 403 Fiber tracking approaches, 285f examples, 284f Figure-ground segregation, 4 Finkels model, 131 Firing types different current injections, 116f examples, 114f Floyd's algorithm, 305 Fluid-dynamical model, 75 FOB, 384 algorithm, 78-80, 79f Forebrain. See Basal forebrain Foreign databases, 99-100 Formal description neuronal networks, 298-299 Formal models, 363...

## Discussion

Stochastic dendritic growth models appear to be successful in describing the shapes of dendritic branching patterns, as shown in section 11.2. and by other authors (92-94). The parameter values, obtained after a process of optimization, are assumed to reflect basic characteristics of the branching process. Emphasis has been given to competitive phenomena as becoming apparant by the size-dependent branching probabilities. We have shown that the competition parameter E significantly...

## Neural Tracing Techniques

The anatomic organization of connections between brain regions, and between the periphery and the central nervous system, are currently studied with increasingly sensitive tracing techniques. These techniques allow a high level of precision and detail in the mapping of topographical patterns (24,43-46) (for a historical review, see 47 ). In our investigations of the cerebro-cerebellar (19,20) and auditory systems (21,22), we have used the neural tracers wheat germ agglutinin-horsesradish...

## Ab

Differential density scatter plots and isorelational surface mapping. (A and B) represent the spatial distribution of cholinergic (dots in panel A) and parvalbumin (dots in panel B) cells from the same brain showed separately. Filled circles mark the high density locations where the density of cholinergic or parvalbumin cells is higher than 15 cells in the unit space (250 x 250 x 50 im). (C) The scatter plots of both cholinergic (red in the color version of this figure) and parvalbumin...

## Claus C Hilgetag PhD

Hilgetag studied Biophysics in Berlin and Neuroscience in Edinburgh, Oxford, Newcastle, and Boston. He is an Assistant Professor of Neuroscience at the newly founded International University Bremen. His research focuses on the organization of cerebral connectivity and architecture as well as the neural mechanisms of spatial attention. Further information can be found at http

## I

Sensory map transformations, 202-204 Imaging toxicology, 387-397 Imaging histology definition, 97 Imaging technique, 13, 385-387 Implementation ArborVitae, 13 construction method, 261-262 Inhibitory interneurons role, 4 Inhibitory synaptic currents, 373 Innervating axons coexistence of stable states, 237f Input-integration-output, 5 Integration method CVODE, 132-133 Integrative aims neuroscientists, 247-248 Interdaughter angle, 41 Interfacing models, 365 Interstitial neurons, 179 Intracellular...

## Introduction

Connectivity defines the role of individual nerve cells, or of distinct neuronal systems, within the global context of neural networks and the brain. Afferents determine the input into a cell or system, and output projections relay the processed information onto selected targets. Perhaps the most intuitive image of the nervous system is that of a network, and many models of information processing in the brain are based upon selected features of neural connectivity, e.g., (1-3). However, the...

## Overview Of The Modelers Workspace

The Modeler's Workspace project is an effort to create a software environment that will make it easier for biologists to develop, use, and share structurally realistic models and, thereby, gain some of the benefits discussed above. As mentioned in the introduction, our goal is to provide the following kinds of facilities in the system 1. Search and retrieval facilities for interacting with multiple databases of models and other information. 2. Facilities for creating, editing and visualizing...

## B

Firing types at different current injections in nA. Each of the 16 CA3 pyramidal cells tested is listed by column and ordered from smallest (left) to largest (right) dendritic area. Each row represents the amount of current injected at the cell's soma. Each entry in the table denotes the qualitative response of a particular cell to a level of current injection. Adapted from (44) with permission.

## Mm

A prototype of the Search pane. The upper left area allows the user to specify the object characteristics to search for the upper right area allows the user to specify which databases should be used in a given search and the bottom half provides a summary of the search results. corresponding to the chosen template. This allows the user to specify the attribute values on which to search. The form is similar to that presented by the Generic Inspector mentioned above. The second region in...

## Structure Of Adult Maps

Receptive fields of V1 neurons are highly selective along a number of feature dimensions of the stimulus. These feature dimensions include position in the visual field, eye of origin (ocular dominance), orientation, direction of movement, spatial frequency of a grating, and disparity. Neurons lying along a line or column orthogonal to the surface of V1 respond in approximately the same way to visual stimuli. However, responses vary in an organized way in the tangential direction, parallel to...

## Conceptual Background

The theoretical framework of this work grew out of methodological studies of prevailing models of the agent in microeconomic, game, and decision theory (1,2). The basic finding was that these models typically presupposed agents with unlimited computational capacities, and more realistic bounded-resource models were then developed. Subsequently, the same approach was applied in computer science, to connectionist models of massively parallel and interconnected computation that were intended to be...

## Original distances

Shepard scatterplot corresponding to Figure 6, for the scaled distances in 2D space (on the y-axis) versus the original dissimilarity distances (on x-axis). While there is a trend for smaller original distances to be represented by smaller reproduced distances, the scaled distances that reproduce the original three data categories overlap widely, indicating an only moderate fit between the original and the scaled configuration. based on a decomposition of the covariance matrix of the...

## Results

The simulated network showed stimulus-correlated propagation of activity from LGN to V1 and all other areas. Histograms, of spike time distributions over 100 runs with random background activity of the LGN model (Fig. 2), confirmed that the firing Fig. 2. Histograms of temporal distributions of simulated action potentials in LGN and all units of the 13 cortical areas to visual stimuli lasting from 0-0.5 s (abscissa shows time in s). Acronyms L23, L4, and L56 refer to supragranular, granular,...

## Rule of

In section 2 of this chapter, we described the rule of 1 2 to empirically determine the appropriate time step for a simulation. A similar approach can be used to determine the appropriate size of compartments (28). We can call it rule of 1 3, since it is based on the division of each existing compartment into three equal pieces (the use of an odd number guarantees that a virtual electrode in the middle of a compartment does not need to be repositioned upon the division). If the numerical...

## References

Forss J, Beeman D, Eichler-West R, Bower JM. The Modeler's Workspace A distributed digital library for neuroscience. Future Generation Computer Systems 1999 16 111-121. 2. Bower JM, Beeman D. The Book of GENESIS Exploring Realistic Neural Models with the GEneral NEural Simulation System, 2nd ed. Springer-Verlag, New York, 1998. 3. Hines M, Carnevale NT. The NEURON simulation environment. Neural Computation 1997 9 1179-1209. 4. Ermentrout GB. XPP-Aut X-windows PhasePlane plus Auto. (http...

## Graph Theoretical Analysis

In what follows, we introduce a number of graph theoretical measures of increasing complexity. We briefly discuss potential neural correlates for each of these measures and indicate how their evaluation may aid in characterizing patterns of anatomical connectivity. 14.3.1. Average Degree of Connectivity A rather crude estimate of the connectivity of a digraph is its average degree of connectivity, that is, the total number of connections present, divided by the total number of connections...

## Robert E Burke md

Burke received his MD degree from the University of Rochester School of Medicine and Dentistry in 1961 and clinical training in Internal Medicine and Neurology at the Massachusetts General Hospital. In 1964 he joined the Spinal Cord Section in the National Institute of Neurological Diseases and Blindness at the National Institutes of Health in Bethesda, MD. He has spent his entire career at NIH and is currently Chief of the Laboratory of Neural Control in NINDS. Dr. Burke's research...

## Computational Implementation

Although standard digital computers are deterministic devices, computer simulations should not be treated literarily. The numerical implementation of mathematical models implies errors in the approximation of the solution, as well as errors in the very representation of numbers in the computer. In this section, we discuss some of these errors, which are intrinsic in numerical methods and computer implementations and are present in addition to any error inherent in the assumptions on which the...

## Miguel A Carreira Perpinan PhD

Carreira-Perpinan has university degrees in Computer Science and Physics (Technical University of Madrid, Spain, 1991) and a PhD in Computer Science (University of Sheffield, UK, 2001), on the use of continuous latent variable models for dimensionality reduction and data reconstruction. His current research interests are statistical pattern recognition and computational neuroscience.

## Imagecombining Microscopy For Data Acquisition

Data entry for the studies reviewed here was made with an image-combining computerized microscope. With this method, data from histological sections, including labeling patterns, were coded as lines and points. The principle of image-combining computerized microscopy was first introduced by Glaser and van der Loos (53,54) and has later been used by numerous investigators (for a review of anatomical data acquisition methods, see 55 ). This system mixes a computer graphical image of digitized...

## David S Lester PhD

Lester is presently the Director of Clinical Technologies at Pharmacia at Peapack, NJ. Previously, he was Senior Science Advisor in the Office of Pharmaceutical Sciences, Center for Drug Evaluation and Research, Food and Drug Administration, Rockville, MD. He earned his PhD from Northwestern University.

## Theoretical Models Coverage And Continuity

The role of activity in shaping cortical maps has usually been modeled via Hebbian learning rules. Such rules can often be interpreted as implementing gradient ascent descent in some objective function, so that the effect of the developmental process is to optimize (at least to some extent) that function. A particularly useful class of objective functions implements a trade-off between two competing tendencies, coverage uniformity (or completeness) and continuity (or similarity). However, even...

## N

For N cities there are possible routes for large N, it is impossible to search them all to find the optimal tour. Therefore, many heuristic algorithms have been investigated, which aim to provide good solutions in reasonable time (for review, see Lawler et al. 47 ). The set of valid tours for a TSP of size N can be represented as the vertices of an N-dimensional hypercube, and most techniques aim to provide good ways of stepping from one vertex to another to gradually...

## Cwl

Absolute latency values, 374 Acetyltransferase, 193 Active models neuronal modeling, 107-109 Activity-independent mechanisms column development, 351-354 AD, 113-120, 172 ADC maps, 276-278, 280f Adjacency Rule, 76 Adult brain weight Conel data, 403f Adult maps visual cortex columnar structures, 338 Adult neuronal morphology model, 9-10 Afferent. See also Thalamocortical afferent (TCA) basal forebrain differential distribution, 184f connectivity patterns, 302f ingrowing construction method,...

## Intense Bursting

Firing types at different current injections in nA. Each of the 16 CA3 pyramidal cells tested is listed by column and ordered from smallest (left) to largest (right) dendritic area. Each row represents the amount of current injected at the cell's soma. Each entry in the table denotes the qualitative response of a particular cell to a level of current injection. Adapted from (44) with permission. ing techniques to detect structure in the data (43). For our analysis, we used Xgobi (64), a...

## L

Labeling techniques, 27 Lamellipodia, 219 Large-scale optimization dendrites and axons, 75-76 Lateral geniculate nucleus (LGN), modeler's workspace, 93 Leaky integrate-and-fire unit, 363 Learning processes, 364 Lemniscal nuclei 3D reconstruction, 210f, 211f pathways, 200-201 LGN, 338, 352 Ligand-activated channel L-Measure, 9, 51, 110 L-Neuron, 8-9 algorithms, 53-55 description, 49 Duke-Southampton format (.swc), 52-57 generation and description, 49-68 graphical formats, 55 modified Hillman...

## Concluding Comments

This chapter has dealt with some approaches to the problem of quantifying the morphology of individual neurons and of ensembles of neurons, using data from cat ventral horn motoneurons. The ability to mimic the statistical properties of cat motoneuron dendrites, viewed in terms of their 2D dendrograms, using a relatively simple growth model based on data extracted from the same data set, provides a parsimonious description of the original data, which separates factors that are determinative...

## The Elastic Net Algorithm

The elastic net algorithm (31) was originally developed as an approximate method for the Traveling Salesman Problem (TSP), a well-known NP-complete combinatorial optimization problem. Here, the objective is to find the shortest distance a salesman can travel to visit a set of N cities in a plane and return to where he started. The key idea is that this problem is analogous to the problem of forming topographic maps in the nervous system, where cities represent input or feature points, and the...

## Conclusions

In this chapter, we reviewed a number of studies that investigated the relationship between neuronal shape and neuronal function. In all cases, variations in shape, even within the same cell class, caused variations in neuronal response. Some of these studies have made an effort to isolate the effect of morphology from variations in physiology (29,34,43,44,68), while other studies have done just the opposite and altered the physiology to overcome differences in morphology (5-8). Now, with the...

## Competition Between Axons In The Refinement Of Neural Circuits

During development, the refinement of neural circuits involves both the formation of new connections and the elimination of existing connections (6,7). Neurons, and other cell types, often are initially innervated by more axons than ultimately maintain into adulthood (7,43). This initial hyperinnervation followed by elimination occurs, for example, in the development of connections between motor neurons and muscle fibers (8,9), where elimination of axons continues until each muscle fiber is...

## H

Hebbian learning rules, 230-232, 339, 346 modified, 231-232 Hierarchical activation indices distribution, 376f Hierarchical analyses, 327-328 Hierarchical cluster tree direct connectivity, 319f global connectivity patterns, 320f High-resolution MRI, 388-390 Hillman, Dean, 7-9 Hillman algorithm modified L-Neuron, 53-55 Hillman's seven fundamental parameters, 28 Hippocampal model, 429 Hippocampal pyramidal cell remodeling, 67f Hippocampus function, 430-431 needle damage, 390 self-location in...

## Contents

1 Computing the Brain and the Computing Brain Giorgio A. 2 Some Approaches to Quantitative Dendritic Morphology Robert E. Burke and William B. 3 Generation and Description of Neuronal Morphology Using L-Neuron Duncan E. Donohue, Ruggero Scorcioni, and Giorgio A. Ascoli 49 4 Optimal-Wiring Models of Neuroanatomy Christopher Cherniak, Zekeria Mokhtarzada, and Uri Nodelman 71 Making Model-Based Studies of the Nervous System More Accessible Michael Hucka, Kavita Shankar, David Beeman, and James M....

## Computer simulation

Diagram representing the phases and components of the mathematical modeling process. considerations are applicable to simulations implemented in other environments, such as Surf-Hippo (6), NEOSIM (7), and Catacomb (8), as well as to simulations written in MatLab (9) or directly in FORTRAN or C++, possibly using libraries such as the Conical Library (10). Most of the aspects examined in this chapter concern the computational implementation of a neurobiological mathematical model (Fig....

## G

Gamma-aminobutyric acid (GABA), 369 Ganglion placement optimization Caenorhabditis elegans, 76-80 Gaussian centers, 346 GenAlg, 78-79, 78f Gene expression models, 353-354 Gene network, 353-354 integrators, 361 GENESIS, 10, 84, 95, 111, 127, 362, 368 Menschik model, 130-132, 130f neural simulator format, 113 web site, 102 Genetic algorithm, 78f Geometry Glial fibrillary acidic protein (GFAP), 384 Global connectivity patterns hierarchical cluster tree, 320f Global distortions, 350-351 Goodhill...

## Benjamin Harrison Landing md

Landing was an Emeritus Professor of Pediatrics and Pathology at USC and served as director of laboratories and chair of pediatric pathology at Children's Hospital Los Angeles from 1959 to 1988. He was one of the founders of Pediatric Pathology and trained more pediatric pathologists than anyone of his generation. His discoveries include defining the structure of skeletal muscle and liver, discovering GM1 gangliosidosis, and overturning the dogma of no mammalian post-natal...

## Contributors

Ascoli, phd Krasnow Institute for Advanced Study and Department of Psychology, George Mason University, Fairfax, VA David Beeman, phd Department of Electrical and Computer Engineering, University of Colorado, Boulder, CO Jan G. Bjaalie, md, phd Institute of Basic Medical Sciences, Department of Anatomy, University of Oslo, Oslo, Norway Sybrand Boer-Iwema, ms Krasnow Institute for Advanced Study, George Mason University, Fairfax, VA James M. Bower, phd Research Imaging Center,...

## O

Object-oriented style advantages, 98-99 Object recognition, 4 Occipital cortex, 421 Ocular dominance EN, 347-348 orientation maps, 350 Ocular dominance column formation, 338 Ocular dominance maps, 352 quantitative measures, 389-394 Onset response latencies, 373, 377 distribution, 375f Optical imaging, 337 Optic nerve stimulating cuff, 353 Optimal cluster arrangements connectivity data, 326f Optimal-wiring models neuroanatomy, 71-80 Optimization analyses, 322-328 Optimization mechanisms, 75-80...

## Au

Comparison between two approaches for the fiber tracking. In this example, an ROI delineated the genu of the corpus callosum, which included 21 pixels. In (A), tracking was initiated from the 21 pixels, resulting in 21 lines to reveal the callosal connections between the frontal lobes. In (B), tracking was initiated from all pixels in the brain, and tracking results that penetrated the ROI were searched. This approach resulted in identifying 1880 pixels that were connected to the ROI,...

## Kavita Shankar PhD

Kavita Shankar received both her BS in Human Biology (1984) and MS in Anatomy (1986) from the All India Institute of Medical Sciences, New Delhi, and a PhD in Craniofacial Biology (1993) from the University of Southern California, Los Angeles. She was the recipient of the NIH-National Research Service Award grant for carrying out her doctoral work. She previously worked in UCLA's Laboratory of Neuroimaging. Her current interests include neuroinformatics and neuroimaging.

## Visualization And Quantitative Analyses Of The Distribution Of Labeled Axons And Cells

Slicing of 3D Reconstructions Traditional anatomic studies are based on observations made in serial sections. The plane of sectioning may, however, lead to incomplete interpretations of topographic organization. In addition to real time rotations of 3D reconstructions and diagrams showing the total projection patterns from several angles of view (Figs. 3 and 4), dynamic subdividing of the complete reconstruction into sections (here referred to as slices) of chosen thickness and...

## Inhomogeneous Distribution Of Chemically Identified Cell Populations

Differential Density 3D Scatter Plot Figure 4A, using a differential density 3D scatter plot (for a brief description of this method, see Appendix 9.9.4.) shows that the density of cholinergic cells is not uniform (see also Fig. 2 in 4 ). Cholinergic cells often form clusters consisting of 3-15 tightly packed cell bodies. The saliency of these clusters, nonetheless, depends on the density threshold setting. For example, when using a relatively low threshold (d > 5 cells per 250 x 250 x...

## The Underlying Architecture

The previous section makes clear that the Modeler's Workspace User Interface is the most essential part of the system. It must provide interfaces not only for interacting with model components, but also with databases and simulation tools. Because model representations, databases, and simulation analysis tools will all change and evolve over time, the User Interface must itself be easily adapted and extended as the needs of users change. We knew from the outset that the success of the project...

## Neurons In Three Dimensions

2D morphological data are relatively tractable for computational modeling, as exemplified by the discussion so far. However, it is considerably more difficult to extend such approaches to neurons as 3D entities (17). The overall shape of neuronal dendritic trees have been analyzed by statistical methods, such as principle components (23,24), and by a Fourier transform technique that can give concise information about the density of branches distributed in 3D space (25). Cullheim and colleagues...

## Representation Of Models And Data

One of the most difficult conceptual issues has been developing a strategy for describing models and their components. The Modeler's Workspace requires a representation language that abstracts away specifics of particular simulators, such as GENESIS, and also provides ways of interacting with existing neuroscience databases on the Internet. Devising such a representation is difficult. The thorniest issue has been balancing the need for specificity in the representation (so that we can develop...

## Anatomical Representation

Equation 3 describes the mathematical model of the spatio-temporal distribution of membrane voltage of a neuron. The numerical approximation of the solution of this problem requires (in analogy to time discretization) space discretization, i.e., the division of the continuous dendritic branches in a finite number of compartments. Each of the state variables in Equation 3 is assumed constant within any one compartment and calculated only in the center of each compartment. Thus, for example,...

## Interacting With Databases

The Modeler's Workspace design supports the ability for users to interact not only with theirs and other users' Workspace Databases, but with databases that were not designed specifically for the Workspace. In this section, we describe how the User Interface component of the system interacts with Workspace Databases and foreign databases. Separating the Workspace Database from the User Interface, and making the former be a stand-alone server, is essential for providing the desired functionality...

## To

The graphs of these gamma curves and corresponding real points (all points that were not of value 1) are shown in Figure 4. The graphs and the linear correlations both indicate the high quality of the gamma fit for these basic parameter data. The measured number of apical trees corresponded to a simple bimodal distribution 22 out of the 24 cells had a single apical tree, while the remaining 2 cells had two apical trees. The corresponding parameter was thus set as a mixture...

## Network Optimization Theory

The theory of NP-completeness emerged around 1972 (11,11a) the key formal concept of a computational problem being NP-complete (nondeterministic polynomial-time complete) is strongly conjectured to be linked with a problem being intrinsically computationally intractable that is, not generally solvable without exhaustive search of all possible solutions. Because the number of possibilities combinatorially explodes as the size of a problem-instance grows, such brute force searches are extremely...

## Michael Hucka PhD

Michael Hucka received his PhD in Computer Science and Engineering in 1998 from the University of Michigan. He has worked in the areas of artificial intelligence, cognitive science, and computational neuroscience. He is currently engaged in developing software to help neurobiologists interact with simulation tools, databases, and other resources.

## Si I

Anterograde axonal tracing of pontine projections from SI in rats (modified from 19 ). (A) Photomicrograph of a frontal section through the center of a BDA injection site placed in the right SI trunk representation under electrophysiological guidance. A bundle of labeled callosal fibers emerge from the injection site. Labeling is also visible in the secondary somatosensory cortex, SII, and thalamus (asterisk). (B) Line drawing of the SI somatotopic map (modified with permission from 9...

## Brief History Of Computers

What will the twentieth century be remembered for in a thousand years from now Here is a guess the emergence of computers1. Apparently, the impact of computers on our society was underestimated at the very beginning. At their birth, computers (literally meaning calculators) were designated for calculating ballistic trajectories in artillery. Here are several definitions of a computer borrowed from modern online dictionaries 1. A machine for performing calculations automatically (WordNet 1.6,...

## Spatial Orientation

A final item of discussion concerns the issue of the spatial orientation of dendrites. While dendrograms capture a great deal of morphological properties and are sufficient to run single-cell electrophysiological simulations, an important component of dendritic morphology is the occupation of space in three dimensions. In fact, this is one of the most important shape characteristics that neuroanatomists intuitively use in morphological classifications. L-Neuron tackles the problem of dendritic...

## Arjen van Ooyen PhD

Arjen van Ooyen is a researcher at the Netherlands Institute for Brain Research. He has a PhD in theoretical neurobiology from the University of Amsterdam. His principal research concerns modeling neural development neurite outgrowth, axon guidance, and axonal competition. Further information can be found on his website at www.anc.ed.ac.uk arjen. Jaap van Pelt received his PhD in Physics in 1978 at the Free University in Amsterdam. His research group, Neurons and Networks, at the Netherlands...

## Ruggero Scorcioni bs

Ruggero Scorcioni is the software engineer of the Computational Neuroanatomy Group at the Krasnow Institute for Advanced Study while studying for a PhD in the School of Computational Science at George Mason University. He graduated in electronic engineering from the University of Modena, Italy.

## Stephen L Senft PhD

Stephen Senft is interested in visualization of brain anatomy and activity at the cellular level. He began his study of Neuroscience - ' with Steve George at Amherst College, and pursued his graduate study of Neuroscience first at the University of Oregon and later at Washington University, with intervening study in Dan Alkon's laboratory at the MBL. He obtained his Ph.D. in the Woolsey laboratory, with an analysis of the ingrowth of thalamic afferents to mouse somatosensory barrel cortex....

## Sybrand Boer Iwema ms

Sybrand Boer-Iwema received an MS in Chemistry from the University of Leiden Netherlands in 2001, spending the last semester of his curriculum as a student intern in the Computational Neuroanatomy Group at the Krasnow Institute for Advanced Study. His interests include neuroscience, computational modeling, and biking. From Computational Neuroanatomy Edited by Giorgio A. Ascoli Humana Press Inc., Totowa, NJ

## Geoffrey J Goodhill PhD

Goodhill has a BSc in Mathematics and Physics from the University of Bristol, an MSc in Artificial Intelligence from the University of Edinburgh, and a PhD in Cognitive Science from the University of Sussex. Following post-doctoral training in computational neuroscience at the University of Edinburgh, Baylor College of Medicine, and the Salk Institute, he joined the faculty of Georgetown University in 1996. He is currently an Associate Professor in the Department of Neuroscience.