Health Articles

426 4.9

W. Bender1M. Albus1H.-J. Möller2 Towards Systemic Theories in Biological Psychiatry Although still rather controversial, empirical data on the neuro- netics, informatics, computational neuroscience or systems sci- biology of schizophrenia have reached a degree of complexity ence. The methodology of systems science permits the modeling that makes it hard to obtain a coherent picture of the malfunc- of complex dynamic nonlinear systems. Such procedures might tions of the brain in schizophrenia. Theoretical neuropsychiatry help us to understand brain functions and the disorders and ac- should therefore use the tools of theoretical sciences like cyber- tions of psychiatric drugs better.
Empirical data and theoretical neuropsychiatry nity state etc. (Leuner and Müller, in this issue p. S15). For ªnatur-alisticº modeling, we do not know enough about the dynamics of Schizophrenia is a mental disorder for which therapeutic im- the dopamine system during sleep-wake cycles, either [42]. Final- provements in antipsychotic medications still are needed ly, the effects of new antipsychotic drugs like aripiprazole which [35,36,39]. Obviously the basic neurobiological mechanisms of acts as a partial dopamine agonist open up new questions with re- this disease are not understood sufficiently (Carlsson, in this is- gard to ªoptimizationº of dopaminergic transmission as basis of sue p. S10). On the other hand, in the last years a huge amount therapeutic effects in schizophrenia (Gründer et al., in this issue of data has been gathered by neurobiological research in the p. S21). These aspects have not yet been clarified by a consistent, form of animal experiments, clinical studies, imaging studies detailed and ªnaturalisticº biologically based theory of dopamine and other methods. Many details are still not clarified as for functions in schizophrenia. Any assumption like the hypothesis or instance it can be discussed controversially if a striatal theory of ªhyperactivity of dopamine transmissionº is a generali- hyperactivity or a prefrontal cortical hypoactivity of dopamine zation which is to some extent contradictory, but possibly not sig- transmission mechanisms is the primary causal mechanism of nificant. In spite of such inconsistencies and uncertainities on the schizophrenia [17,38]. One point of dissension is the conception empirical level, it seems useful to integrate current knowledge of of ªactivityº, as it can be understood as stimulus-bound ªreactiv- the mechanisms of schizophrenia (Emrich et al., in this issue p.
ityº of systems, i.e. their dynamic properties, and can mean tonic S52). Within this context, an enormous amount of information (sustained) and phasic (transient) activity [1,19]. Furthermore, must be compiled in order to construct a picture of this disease.
on the level of the molecular mechanisms of synaptic action, noconsistent picture can be drawn (Leuner and Müller, in this issuep. S15). One example is the question of the relevance of the local Philosophical aspects D1 : D2 ratio of receptors and their effects on local neuronal cir-cuits ([59]; Winterer, in this issue p. S68). The dopamine D 2 type From a philosophical point of view [6], we are in a position in receptors also demonstrate a high degree of diversity occurring which, pictorially speaking, we have broken the entire brain, as as a long and a short subtype, they also have presynaptic, extra- an operationally closed organ, down into it's molecular pieces, synaptic and postsynaptic locations, exert a high and a low affi- but now do not know how to put things together. It is hard to un- 1 District Hospital Haar, Haar, Germany 2 Psychiatric University Hospital, Munich, Germany PD Dr. Dr. Dr. Felix Tretter ´ District Hospital Haar ´ Ringstr. 9 ´ D-85529 Haar/Munich ´ Germany ´ Pharmacopsychiatry 2006; 39 Suppl 1: S4±S9  Georg Thieme Verlag KG Stuttgart ´ New York

derstand a complex mental disease like schizophrenia on the ba- Systems science and similar disciplines sis of disorders of biomolecules. Above all, the effects of suchmolecules on the electrical properties of cells are not yet fully Viewing a subject as a complex of interrelated parts is a basic fea- understood [2]. This reflects the epistemological problem of re- ture of systems science. Integrating different levels of resolution ductionistic top-down analysis with regard to bottom-up expla- of research perspective (micro-/meso-/macro-perspective) into nations resulting in the part-whole problem (Fig.1). To give an an interrelated view is considered essential in the field of sys- example: The properties of oxygen and hydrogen cannot ªex- tems science. Systems science is understood as the ensemble of plainº the behavior of water (emergence-problem).
concepts, methods, theories, models and paradigms used in sys-tems research ([55]. Systems analysis, within the scope of sys- When attempting to construct models, one must first consider tems science, not only refers to a black box-type of input-output the fact that ªrealityº is, at least partially, also a ªconstructionº analysis but also explicitly considers the contextual conditions of [14,18]. Therefore any theory is not ªabstractº but only a higher the system under study (Fig. 2). The system is seen as a struc- order construction. Conversely, according to Kurt Lewin, theories tured entity within a structured environment.
are very ªpracticalº, as they can help us reduce empirical com- plexity to the essentials.
With regard to these methodological issues, the question alsoarises of which academic discipline would offer a good partner- It should be borne in mind that theories and models need not be ship for theoretical neuropsychiatry. Within the field of cyber- ªexplanationsº of observed phenomena but can also be useful as netics [4,58], the term ªneurocyberneticsº was used only seldom to characterize a specialization within this field devoted to brainresearch. This term could not be established. Subsequently, the Furthermore, with regard to the complexity of known brain term ªneuroinformaticsº was coined, which is still in use [3].
mechanisms, it should be kept in mind that models are always The study of ªartificial neural networksº was successfully devel- selections from observed reality. It is essential to keep this in oped in this field ([22,27,34,47,51]; Hoffmann and McGlashan, mind, as the brain is a system with billions of neuronal feedback in this issue p. S54 and Deco, in this issue p. S65). In theoretical loops. It is thus hard to ªlocalizeº a function (ªcircular causalityº).
brain research, the term ªcomputational neuroscienceº is nowused most [53]. Dayan and Abbott [10] also use the term ªtheore- There is also a large conceptual gap between empirical research tical neuroscienceº (Dayan and Williams, in this issue p. S50).
and theory which ± especially in brain research ± must bebridged by an interdisciplinary formal language closely related These disciplines are more technical and do not pay so much at- to mathematics. In this respect, ªsystems scienceº, or ªcomputa- tention to the basic methodological problems of modeling the tional neuroscienceº, offers a conceptual and methodological ba- brain, as understanding the brain in principle requires a tool for sis for integrating the various data within a sophisticated system cognitive management of complexity. Some authors use the term ªsynergeticsº to characterize the study of dynamic order [20,21], others search for a ªcomplexity scienceº [31,32]. Traditionally,these questions are discussed in the context of systems science.
It should also be mentioned that, in molecular biology, some re- searchers such as Kitano [24], are convinced that a systemic per- tionistic approach ± spective is important nowadays ± experiments and theory build- top-down process of empirical and ex- ing must be designed in such a way that the interrelationships and interactions of the various molecules can be considered.
roscience und the This new field, which can be referred to the conceptual horizons problem to explain of the gene regulation model of Jacob [23] and Monod [37] is Fig. 2 The multi-level perspective' of systemic thinking. When a sys- tem is being analysed,consideration is also given to the fact that the system is part of a supra-system and the elements are composed of sev- eral components that also are systems (subsystems).
Bender W et al. Towards Systemic Theories ¼ Pharmacopsychiatry 2006; 39 Suppl 1: S4 ±S9 called ªsystems biologyº [25] or ªcomputational systems biolo- As biological sciences are considered increasingly important for gyº [26]. Some researchers have recently created a field called understanding mental diseases, psychiatrists also have to refer ªsystems neurobiologyº for which introductory textbooks are to those theoretical sciences that are used in brain research.
not yet available.
What is important is, that building models for explanation pur-poses necessitates the use of computer simulations to describe Theoretical neuropsychiatry faces a special problem, however: In the course of activity of all the significant components of the cir- computational science, neuroinformatics, systems science or cy- cuitry involved in schizophrenia. The acquisition of such model- bernetics and similar academic disciplines, the analysis of elec- building skills appears fruitful and special training in this field tric signals dominates. The analysis of chemical signalling is un- might be useful for psychiatrists. Several institutes such as the common. This implies difficulties for theoretical concepts.
Volen Center for Complex Systems in Waltham or the Salk Insti-tute in San Diego have thus already produced models concerned Regardless of which concept might be ªbestº for a deeper under- with psychiatric diseases like those of Wang or of Sejnowski, standing of the brain functions, we will give the name systems Seamans and Durstewitz on the function of dopamine receptors science' to an approach which basically regards its field of analy- with regard to working memory and their relevance for schizo- sis as a system [40,43, 55].
phrenia ([11,12,57]; Wang, in this issue p. S80 and Durstewitz,in this issue p. S72).
The brain, neuroscience, psychiatry and systems science It should only be mentioned here that the mind-brain problemmust also be considered. In this context, we think that some The brain is a complex, dynamic, self-organizing, multi-level sys- sort of ªtranslationº of psychological/psychopathological cate- tem. This complexity problem, resulting from the number of gories into cybernetic language (terms like working memory, fil- neurons (approx. 1011) and the number of connections (approx.
ter operations, dynamics, bifurcation, attractor, networks etc.) 1015), must be considered much more in neuropsychiatric re- might be useful to bridge this gap ([41]; Dayan and Williams, in search [5,31±33]. One consequence of these figures is that, on this issue p. S50 and Hoffman and McGlashan, in this issue p.
average, after about four connection steps, a neuron already re- S54). However, we will focus on modeling strategies.
ceives its feedback. The brain is thus a system of billions of neu-ronal circuits. For this reason, brain research set up contacts withcybernetics and systems science, which were emerging in the Neurochemical interaction matrix 1950ies and 1960ies, at a very early stage. Parallel to these theo-retical efforts, technical development in the field of electrical en- The identification of local brain structures exhibiting dysfunc- gineering also brought technical devices like amplifiers that al- tions in psychiatric diseases usually focuses on various brain lowed electrical signals from single nerve cells to be measured.
areas like the striatum or the prefrontal cortex. However, experi- With regard to single unit activity, the ªbinding problemº is dis- ences with the effects of psychopharmacological drugs and illicit cussed in neurobiology, which concerns to the question of how drugs point to an assignment of clinical phenomena to global the activity of a population of single neurons is correlated to a neurochemically defined cerebral networks. In the context of cognitive event [49,50]. This problem means that we do not this idea, a model of the interactions between these systems can know what a single action potential or a brief discharge pattern be constructed. The form of interactions in healthy state and in (burst) of a pyramidal cell in the cortex means on the behavioral state of disorder and the action of drugs should be represented level. On the other hand, the EEG patterns, as macro-signals, are mostly too unspecific to be correlated to a particular behavior(Gallinat and Heinz, in this issue p. S76). This problem was dis- With regard to schizophrenia, a hyperactivity of the inhibitory cussed in detail in theoretical neurobiology by W. Freeman, dopamine D2 transmission system [7] and/or of the activating who recommends measuring the ªmesoscopicº aggregates of serotonin 5-HT2 transmission system [13], hypoactivity of the the brain and building ªmesoscopic modelsº of the brain func- excitatory glutamate NMDA transmission system [48] and hy- tions [15]. These models are intended to describe the activity of poactivity of the inhibitory GABA transmission system [28] are assemblies of neurons. To this extent, the analysis of field poten- discussed alternatively and also with regard to their interdepen- tials and the degree of discharge coherence might be a route to dencies [8]. Taking into account the fact that these systems, and understanding cognitive processes such as pattern recognition.
also excitatory acetylcholine and excitatory norepinephrine The synchronous discharges of populations of neurons with transmission, are also relevant to understand the symptomatol- 30±70 Hz (gamma frequency) might be relevant here, especially ogy of depression and addiction, it seems useful to design a basic in cognitive disorders like schizophrenia (Wang, in this issue p.
structural model representing all these six systems and the pos- sible fifteen interrelations (or 30 interactions; [9]). We call thisintegrative framework the ªneurochemical interaction matrixº In biological psychiatry, however, the focus is on processes in [56] and it is depicted here by a diagram (Fig. 3). Important inter- which biomolecules are involved. With regard to the molecular ferences with this network by various drugs can also be repre- domain of present psychiatric research, the micro-/macro-prob- sented in this model (Fig. 4). It should be mentioned here that lem of explanations, mentioned above, is present, as the way in the matrix concept can be the starting point for formal modeling which molecules work together in the entire brain must be ex- and also allows visualization. Visualizations are very useful for plained. However, there is not enough study of the interaction representing the basic conceptualization of a model and help us between different transmitter systems.
to understand the dynamics of complex systems intuitively.
Bender W et al. Towards Systemic Theories¼ Pharmacopsychiatry 2006; 39 Suppl 1: S4±S9

Fig. 3 The ªneuro- As a next step, this structure-oriented graphical model should be chemical interaction transformed into a formal functional model reflecting the dy- matrixº as a heuris- namics of the various processes. However, a formal model, con- tic schema [56]. Six sisting, for instance, of a system of six coupled nonlinear differ- neurochemical sys- ential equations, cannot be solved mathematically without using repinephrine (NE), numerical methods. For such solutions, the quantification of the acetylcholine (ACh), couplings between the components of the system, namely the synapse dynamics, must be known. This is currently not the glutamate (Glu),do- case. Therefore, only partial modeling is possible ([44]; Tretter and Scherer, in this issue p. S26). Here, as an example, we focus ric acid (GABA),and on the basic circuitry model from Arvid Carlsson ([7]; Carlsson, their 15 interrela- in this issue p. S10) and select the relevant components from tions are depicted, the entire system (Fig. 5). However, the behavior of three recipro- some of them expli- cally interconnected elements is already impossible to foresee by citly. Annotation: Ar- rows represent acti- the imagination alone. For exploratory purposes, this model can vations,lines with transoms stand for inhibitions. Scored lines without therefore be modeled and tested by computer simulations (Tret- arrows or transoms at the end of the line stand for relations not inter- ter and Scherer, in this issue p. S26) and formal analyses preted here.
(Schwegler, in this issue p. S43). Simulations performed by suchmodels help identify relevant questions for future empirical re-search. It should be kept in mind here, that the value of a model Even a first glance at the two matrices reveals the high level of thus lies in the purpose of its construction, and it is therefore not complexity of the reactions of the neurochemical network. How- necessary to represent all the details of reality! ever, it must also be kept in mind that every system has severalsubsystems with regard to the various receptor subtypes [9]. For As appropriate empirical data are not available, computational instance, the dopamine system might have excitatory D1 recep- modeling of the entire neurochemical matrix is not very useful tors mainly on glutamate neurons and inhibitory D2 receptors at present. However, the dynamics of the entire system, for heur- mainly on GABA neurons [45,52]. With regard to these connec- istic purposes, can be represented in the framework of a mobile, tions, a high level of dopamine release would simultaneously which could allow for better understanding of complex dynam- highly activate D1 receptors and D2 receptors, thus pushing very strongly an accelerator and a brake, simultaneously. But D1and D2 receptors probably demonstrate differences in sensitivityresulting in different temporal dynamics (Durstewitz, in this is- The neurochemical mobile ± a heuristic scheme for complex sue p. S72). Macroanatomical details should also be considered in this scheme, as most authors think that the cortical dopaminefunction is reduced, when the striatal function is elevated (Grün- The concept of a ªneurochemical mobileº was proposed by Jür- der et al., in this issue p. S21 and Winterer, in this issue p. S68).
gen Fritze [16]. It is an integrative framework and a metaphorical Also this aspect could be integrated in this matrix.
concept representing complex nonlinear dynamics. This mobileconcept was further developed as a heuristic tool for clinical ob- Fig. 4 Influences on the operation of the neurochemical interaction matrix [56]. Sur- vey of actions of psychopharmacological drugs and illicit drugs on various system components. Contact with a system com- ponent can lead to changes in another sys- tem component. Legend: s. Fig. 3.
Bender W et al. Towards Systemic Theories ¼ Pharmacopsychiatry 2006; 39 Suppl 1: S4 ±S9

Fig. 7 The neurochemical mobile in the state of schizophrenia and the effects of antipsychotic drugs ± relative dominance of dopamine and serotonin and relative hypoactivity of glutamate and the equili- Fig. 5 Structure of the interaction of systems crucial for schizophrenia brating effects of antipsychotic medications,and additionally,of ben- with regard to the basic model of Carlsson [7] and as a selection from zodiazepines (adapted from [56]).
the entire neurochemical matrix. The system,consisting of two feed- back loops,in many constellations of parameters tends to oscillate or show chaotic discharge patterns [29,54]. Legend: Stippled lines: weak Fig. 8 D2 receptor- action,bold lines: strong influence,arrows: excitation,transoms: inhi- based transmission bition; 5-HT not considered here.
globally dominating in schizophrenia,as servations that helps to describe the neurochemical basis of sev- D2 receptor blockers eral psychiatric disorders and their pharmacological therapy reduce the psychotic [56]. The mobile consists of five linked, but oscillating, scales state and support where each of the six scale pans represents a transmitter system re-establishment of a functional equili- with mutually antagonistic effects. The size of the scale pan indi- brium (stippled line).
cates the functional weight of the respective system within the overall system (Fig. 6). In this conceptual framework, activatingand inhibiting (or: antagonizing) substances are located in oppo-site positions on the respective scale. Furthermore, if a scale pan new scale, thus representing subsystems. In order to characterize is, so to speak, full of transmitters', it changes the relations to the some relations in schizophrenia, this is shown with dopamine, other scale pans. Thus, in the case of schizophrenia, the hyperac- depicting the excitatory D1 receptor and inhibitory D2 receptor tivity of serotonin and/or dopamine and/or the hypoactivity of effects (Fig. 8) (Winterer, in this issue p. S68). Thus, consideration glutamate and/or GABA can be represented in a disbalanced con- of receptor subtypes and topographical differentiations can, in figuration of the mobile (Fig. 7). Consequently, one can ªexplainº principle, also be done within the framework of the ªneurochem- the therapeutic effects of antipsychotic drugs and even of benzo- ical mobileº.
diazepines (Fig. 7). For more detailed consideration, each scalepan, as a component of the mobile, can be subdivided into a It should be mentioned that the neurochemical mobile is, con-ceptually, also in line with the mathematical theory of ªdyna-mical diseasesº ([30]; an der Heiden, in this issue p. S36). Themobile indicates that the brain is a nonlinear neurochemical os-cillator. However, when developing a mathematical model thesame restrictions as mentioned for the neurochemical matrixapply.
The aim of this paper was to propose that the complexity of thebrain and the complexity of data of biological psychiatry demanda sophisticated systematic approach towards building modelswhich allow a functional understanding of the brain in psychia-tric diseases. There is much evidence of interconnected neuro- Fig. 6 The ªneurochemical mobileº in a balanced state (from [56]).
chemical feedback and feedforward loops. The integration of Bender W et al. Towards Systemic Theories¼ Pharmacopsychiatry 2006; 39 Suppl 1: S4±S9 concepts, methods and models of systems science or computa- 30 Mackey MC, an der Heiden U. Dynamical diseases and bifurcations.
tional neuroscience might therefore be fruitful for understand- Funkt Biol Med 1982; 1: 156±164 ing the interplay of brain chemistry in psychiatric diseases. The Mainzer K. Thinking in Complexity. The Computational Dynamics of Matter, Mind, and Mankind. New York: Springer,; 4th edition 2004 concept of a neurochemical interaction matrix and the metapho- 32 Mainzer K. Symmetry and Complexity. The Spirit and Beauty of Non- rical ªneurochemical mobileº might be steps towards mathema- linear Science. Singapore: World Scientific Publ, 2005 tical and computational modeling of the neurochemistry of the 33 Mainzer K, Müller A, Saltzer WG. From Simplicity to Complexity: In- 34 McCelland JD, Rumelhart DE and the PDP Research Group. Parallel dis- tributed processing: Explorations in the microstructure of cognition.; Vol 2 Cambridge (MA): MIT Press, 1986 35 Möller HJ, Deister A. Schizophrenie. In: Möller HJ, Laux G, Kapfham- mer HP (Hrsg). Psychiatrie und Psychotherapie Berlin: Springer, 1 Abi-Dargham A, Rodenhiser J, Printz D, Zea-Ponce Y, Gil R, Kegeles LS, 2003: p. 1051±1122 Weiss R, Cooper TB, Mann JJ, Van Heertum RL, Gorman JM, Laruelle M.
36 Möller HJ, Müller WE, Volz HP. Psychopharmakotherapie. Stuttgart: Increased baseline occupancy of D2 receptors by dopamine in schizo- phrenia. Proc Natl Acad Sci USA 2000 Jul 5; 97(14): 8104±9 37 Monod J. Chance and Necessity. New York: Random House, 1971 2 Abbott L, Regehr WG. Synaptic computation. Nature 2004; 431(7010): 38 Moore H, West AR, Grace AA. The regulation of forebrain dopamine transmission: relevance to pathophysiology and psychopathology.
Arbib MA, Grethe JS. Computing the Brain: A Guide to Neuroinformat- Biol Psych 1999; 46: 40±55 ics (hardcover). Academic press, San Diego: 2001 39 Müller WE, Eckert A. Psychopharmakotherapie ± pharmakologische 4 Ashby Ross W. Introduction to cybernetics. London: Methuen, 1964 Grundlagen. In: Möller HJ, Laux G, Kapfhammer HP (eds). Psychiatrie 5 Braitenberg V, Schüz A. Cortex: Statistics and Geometry of Neuronal und Psychotherapie Berlin: Springer, 2003: p. 513±564 Connectivity. Heidelberg: Springer, 1998 40 Ossimitz G. Entwicklung des systemischen Denkens. München: Profil, 6 Bunge M. Philosophy of Science.; 2 Volumes New Brunswick, London: Transaction Publishers, 1998 41 Paulus MP, Braff DL. Chaos and schizophrenia ± does the method fit 7 Carlsson A. The current status of the dopamine hypothesis of schizo- the madness?.Biol Psych 2003; 53: 3±11 phrenia. Neuropsychopharmacology 1988; 1: 179±186 42 Rao ML, Gross G, Halaris A, Huber G, Marler M, Strebel B, Bräunig P, 8 Carlsson A, Waters N, Holm-Waters S, Tedroff J, Nilsson M, Carlsson Klosterkötter J. Circadiane Rhythmik von Catecholaminen, Melatonin ML. Interactions between monoamines, glutamate, and GABA in schi- und hypophysären Hormonen bei schizophrenen Patienten. In: Bau- zophrenia: new evidence. Annu Rev Pharmacol Toxicol 2001; 41: mann, P. (ed). Biologische Psychiatrie der Gegenwart Berlin: Springer, 9 Cooper JR, Bloom FE, Roth RH. The biochemical basis of neuropharma- 43 Richmond B. An introduction to systems thinking. Hanover/USA: High cology. New York: Oxford Univ Press, 2002 Performance Systems, 2001 10 Dayan P, Abbott L. Theoretical neuroscience. Cambridge: MIT Press 44 Scherer J, Eberle E, Tretter F. A basic mathematical model of the dopa- mine synapse. In: Trappl R (ed). Cybernetics and Systems; Volume I 11 Durstewitz D, Seamans JK. The computational role of dopamine D1 re- Austrian Society for Cybernetic Studies, Vienna: 2002: p. 335±328 ceptors in working memory. Neural Netw 2002; 15: 561±572 45 Seamans JK, Gorelova N, Durstewitz D, Yang CR. Bidirectional dopa- 12 Durstewitz D, Seamans JK, Sejnowski TJ. Neurocomputational models mine modulation of GABAergic inhibition in prefrontal cortical pyra- of working memory. Nat Neurosci 2000; 3 Suppl: 1184±1191 midal neurons. J Neurosci 2001 May 15; 15;21(10): 3628 ±38 13 Ereshevsky L, Miller A. Serotonin mechanisms in the pathophysiology 46 Shepherd GM. Neurobiology. Oxford: Oxford Univ. Press, 1994 of schizophrenia. Neuroscience, 1999 Scientific connexions 47 Shortlife EH, Perreault LE, Wiederhold G, Fagan LM(eds). Medical In- 14 von Foerster H. Observing Systems. Seaside, Californien: Intersystems formatics: Computer Applications in Health Care and Biomedicine Publications, 1982 (Health Informatics). Berlin: Springer, 2000 15 Freeman WJ. Neurodynamics. Berlin: Springer, 2000 48 Simpson MD, Slater P, Royston MC, Deakin JF. Regionally selective def- 16 Fritze J. Einführung in die biologische Psychiatrie. Stuttgart: Fischer, icits in uptake sites for glutamate and gama-aminobutyric acid in the basal ganglia in schizophrenia. Psych Res 1992; 42(3): 273±82 17 Gao WJ, Goldman-Rakic PS. Selective modulation of excitatory and in- 49 Singer W. Search for Coherence: a basic principle of cortical self-orga- hibitory microcircuits by dopamine. PNAS 2003; 100(5): 2836±2841 nization. Neuroscience 1990; 1: 1±26 18 Glasersfeld Ev. Radical constructivism: A way of knowing and learn- 50 Singer W. Neuronal synchrony: a versatile code for the definition of ing. London: Falmer Press, 1995 relations. Neuron 1999; 24: 49±65 19 Grace AA, Bunney BS. Electrophysiological Properties of Midbrain Do- 51 Spitzer M. The mind within the net. Cambridge, Mass.: MIT Press,1999 pamine Neurons: Psychopharmacology ± The Fourth Generation of (1996:German edition) Progress. Neuropsychopharmacology Journal, 2000 52 Trantham-Davidson H, Neely LC, Seamans JK. Mechanisms underlying 20 Haken H. Synergetics. Berlin: Springer, 1983 differential D1 versus D2 dopamine receptor regulation of inhibition 21 Haken H. Brain dynamics. Berlin: Springer, 2002 in prefrontal cortex. J Neuroscience 2004; 24(47): 10 652 ±10 659 22 Haykin S. Neural Networks: A Comprehensive Foundation (2nd Edi- 53 Trappenberg T. Fundamentals of Computational Neuroscience. New tion). Upper Saddle River, NJ, USA: Prentice Hall PTR, 1998 York: Oxford Univ. Press, 2002 23 Jacob F. Logic of Life: A History of Heredity. NY: Pantheon, 1973 54 Tretter F. Perspektiven der mathematischen Systemtheorie in der bio- 24 Kitano H (ed.). Foundations of Systems Biology. Cambridge (US): MIT logischen Psychiatrie. Krankenhauspsychiatrie 2004; 15: 77±84 55 Tretter F. Systemtheorie im klinischen Kontext. Lengerich: Pabst, 2005 25 Kitano H. Systems Biology: a brief overview. Science 2002; 295: 56 Tretter F, Albus M. Einführung in die Psychopharmakotherapie.
Thieme, Stuttgart: 2004 26 Kitano H. Computational Systems Biology. Nature 2002; 420: 206± 57 Wang XJ, Tegner J, Constandinidis C, Goldman-Rakic PS. Division of la- bor among distinct subtypes of inhibitory neurons in a cortical micro- 27 Kohonen T. Self-Organizing Maps. Berlin: Springer, 2001 circuit of working memory. PNAS 2004; 101(5): 1368±1373 28 Lewis DA, Pierri JN, Volk DW, Melchitzky DS, Woo TUW. Altered GABA 58 Wiener N. Cybernetics or control and communication in the animal neurotransmission and prefrontal cortical dysfunction in schizophre- and the machine. Cambridge: MIT Press, 1948 nia. Biol Psych 1999; 46: 616±626 59 Winterer G, Weinberger DR. Genes, dopamine and cortical signal-to- 29 Mackey MC, an der Heiden U. The dynamics of recurrent inhibition.
noise ratio in schizophrenia. Trends in Neurosciences 2004; V 27: 11 J Math Biol 1984; 19: 211±225 Bender W et al. Towards Systemic Theories ¼ Pharmacopsychiatry 2006; 39 Suppl 1: S4 ±S9

Source: http://geza.kzoo.edu/~erdi/skread/bender06.pdf