ScienceWeek

NEUROBIOLOGY: ON HYBRID NEURAL NETWORKS

The following points are made by Astrid A. Prinz (Current Biology 2004 14:R661):

1) Faced with the complexities of brain activity, some neuroscientists are turning to hybrid networks -- neural networks consisting of living nerve cells interacting with model neurons -- to help them understand how neural and synaptic properties shape the electrical activity of neural circuits [1,2] or to validate models of neurons in a network setting [3-5].

2) Neural activity is fundamentally complex. Neurons in networks with many synapses and feedback loops constantly receive inputs from other neurons, integrate them and generate electrical activity patterns in response. Network activity is thus shaped by interactions between the non-linear electrical properties of neurons and synapses. These complex interactions allow neural circuits to process information, support cognitive functions, and control behavior.

3) Cellular electrophysiology experiments do not always acknowledge this complexity. Much of our understanding of neural circuits at the small network level relies on highly reductionist experiments. We characterize the response of isolated neurons to simple stimuli such as current injections or voltage steps, or we measure the signal transmitted through a single synapse. These time-proven experimental approaches are essential for our understanding of the building blocks of neural networks, but their reductionist nature raises the question whether we are missing something by probing a complex system with simple perturbations.

4) To examine the behavior of neurons and small networks under more realistic conditions, researchers in the early 1990s began to study the interactions of living nerve cells with model neurons in hybrid networks. Such connections between living and model neurons combine physiological realism with complete control over the neural and synaptic properties of the artificial network components. Hybrid networks thus create an interface between experimental and modeling studies, combining the best of both worlds.

5) The model neurons and synapses in hybrid networks can be digital or analog. In hybrid networks with digital components, a technique called the "dynamic clamp" is used to monitor the membrane potential of living neurons, to numerically simulate model neurons and synapses on a computer, and to inject synaptic currents into living neurons in real-time, as if the living neurons were synaptically connected to the model neurons. Alternatively, the dynamic clamp can be used to insert artificial membrane conductances into living neurons embedded in a network, thus exploring the role of intrinsic conductances in shaping network output.

6) In hybrid networks with analog model neurons and synapses, a specially designed electronic circuit constitutes the artificial part of the network. Such hardware model neurons and synapses are connected to living circuits through electrodes, creating a hybrid circuit that consists of a biological part and a dedicated silicon chip.

References (abridged):

1. Sorensen, M.X DeWeerth, S.X Cymbalyuk, G. and Calabrese, R.L. (2004). Using a hybrid neural system to reveal regulation of neuronal network activity by an intrinsic current. J. Neurosci. 24, 5427-5438

2. Nowotny, T.X Zhigulin, V.P.X Selverston, A.I.X Abarbanel, H.D.I. and Rabinovich, M.I. (2003). Enhancement of synchronization in a hybrid neural circuit by spike-timing dependent plasticity. J. Neurosci. 23, 9776-9785

3. Aliaga, J.X Busca, N.X Minces, V.X Mindlin, G.B.X Pando, B.X Salles, A. and Sczcupak, L. (2003). Electronic neuron within a ganglion of a leech (Hirudo medicinalis). Phys. Rev. E 67, Article #061915

4. Pinto, R.D.X Varona, P.X Volkovskii, A.R.X Szucs, A.X Abarbanel, H.D.I. and Rabinovich, M.I. (2000). Synchronous behavior of two coupled electronic neurons. Phys. Rev. E 62, 2644-2656

5. Szucs, A.X Varona, P.X Volkovskii, A.R.X Abarbanel, H.D.I.X Rabinovich, M.I. and Selverston, A.I. (2000). Interacting biological and electronic neurons generate realistic oscillatory rhythms. Neuroreport 11, 563-569

Current Biology http://www.current-biology.com

--------------------------------

Related Material:

NEURAL NETWORKS AND ETHOLOGY

Notes by ScienceWeek:

The tungara frog (Physalaemus pustulosus), also known as the "mud-puddle frog", is common in the rain forest of Costa Rica. This frog has irregular bumpy skin, short legs, and a squat toad-like body. The animal makes a distinctive "ricochet" call suprisingly loud for its small size (an inch or so in length).

The following points are made by S.M. Phelps et al (Proc. Nat. Acad. Sci. 2001 98:13161):

1) Early ethologists reported that animal signals evolved toward simplicity, specificity, and salience -- hallmarks of the minimal stimulus, the "sign stimulus", that was required to evoke a response from its receiver. This observation suggested that signal form was being shaped by the perceptual mechanisms of the receiver, a view that has been rekindled by recent work in sexual selection and sensory ecology. The renewed interest in proximate causes of behaviors has prompted a number of workers to return to the methodical titration of receiver decision mechanisms used in classic ethology; more recently, the resulting generalization gradients (or preference functions, as they are known in mate choice) are thought to strongly affect the fitness of individual signalers. Although the shapes of generalization gradients are presumed to influence signal evolution, and to be of interest in their own right, few ethologists have addressed the forces that determine these shapes.

2) A number of groups have begun to use artificial neural network models to investigate the evolution of perceptual mechanisms. Because neural network models distribute the representation of a signal across many "neurons", these models often generalize as an automatic result of training, making them useful tools for the exploration of preference functions. In recent studies, the authors evolved artificial neural networks along distinct evolutionary trajectories and found that their emergent responses to novel signals were strongly shaped by their selection histories. Moreover, those networks with a history approximating that of our focal species, the tungara frog, were better at reproducing female responses to test stimuli. These findings and others indicate that female tungara frogs may exhibit preferences that are remnants of past selection for species recognition. They do not, however, suggest what form vestigial preferences might take.

Proc. Nat. Acad. Sci. http://www.pnas.org

--------------------------------

Related Material:

NEUROBIOLOGY: ON NEURONAL NETWORKS

The following points are made by S.B. Laughlin and T.J. Sejnowski (Science 2003 301:1870):

1) Neuronal networks have been extensively studied as computational systems, but they also serve as communications networks in transferring large amounts of information between brain areas. Recent work suggests that their structure and function are governed by basic principles of resource allocation and constraint minimization, and that some of these principles are shared with human-made electronic devices and communications networks.

2) The discovery that neuronal networks follow simple design rules resembling those found in other networks is striking because nervous systems have many unique properties. To generate complicated patterns of behavior, nervous systems have evolved prodigious abilities to process information. Evolution has made use of the rich molecular repertoire, versatility, and adaptability of cells. Neurons can receive and deliver signals at up to 10^(5) synapses and can combine and process synaptic inputs, both linearly and nonlinearly, to implement a rich repertoire of operations that process information (1).

3) Neurons can also establish and change their connections and vary their signaling properties according to a variety of rules. Because many of these changes are driven by spatial and temporal patterns of neural signals, neuronal networks can adapt to circumstances, self-assemble, autocalibrate, and store information by changing their properties according to experience.

4) The simple design rules improve efficiency by reducing (and in some cases minimizing) the resources required to implement a given task. It should come as no surprise that brains have evolved to operate efficiently. Economy and efficiency are guiding principles in physiology that explain, for example, the way in which the lungs, the circulation, and the mitochondria are matched and coregulated to supply energy to muscles (2).

5) Just like the wires connecting components in electronic chips, the connections between neurons occupy a substantial fraction of the total volume, and the wires (axons and dendrites) are expensive to operate because they dissipate energy during signaling. Nature has an important advantage over electronic circuits because components are connected by wires in three-dimensional (3D) space, whereas even the most advanced VLSI (very large scale integration) microprocessor chips use a small number of layers of planar wiring. [A recently produced chip with 174 million transistors has seven layers (3).] Does 3D wiring explain why the volume fraction of wiring in the brain (40 to 60%) is lower than in chips (up to 90%)? In chips, the components are arranged to reduce the total length of wiring. This same design principle has been established in the nematode worm Caenorhabditis elegans, which has 302 neurons arranged in 11 clusters called ganglia. An exhaustive search of alternative ganglion placements shows that the layout of ganglia minimizes wire length (4).

6) Cortical projections in the early sensory processing areas are topographically organized. This is a hallmark of the six-layer neocortex, in contrast to the more diffuse projections in older three-layer structures such as the olfactory cortex and the hippocampus. In the primary visual cortex, for example, neighboring regions of the visual field are represented by neighboring neurons in the cortex. Connectivity is much higher between neurons separated by less than 1 mm than between neurons farther apart, reflecting the need for rapid, local processing within a cortical column -- an arrangement that minimizes wire length. Because cortical neurons have elaborately branched dendritic trees (which serve as input regions) and axonal trees (which project the output to other neurons), it is also possible to predict the optimal geometric patterns of connectivity (5), including the optimal ratios of axonal to dendritic arbor volumes. These conclusions were anticipated nearly 100 years ago by the great neuroanatomist Ramon y Cajal (1852-1934).

7) In summary: Brains perform with remarkable efficiency, are capable of prodigious computation, and are marvels of communication. We are beginning to understand some of the geometric, biophysical, and energy constraints that have governed the evolution of cortical networks. To operate efficiently within these constraints, nature has optimized the structure and function of cortical networks with design principles similar to those used in electronic networks. The brain also exploits the adaptability of biological systems to reconfigure in response to changing needs.(5)

References (abridged):

1. C. Koch, Biophysics of Computation: Information Processing in Single Neurons (Oxford Univ. Press, New York, 1999)

2. E. R. Weibel, Symmorphosis: On Form and Function in Shaping Life (Harvard Univ. Press, Cambridge, MA, 2000)

3. J. D. Warnock et al., IBM J. Res. Dev. 46, 27 (2002)

4. C. Cherniak, J. Neurosci. 14, 2408 (1994)

5. G. Mitchison, Proc. R. Soc. London Ser. B 245, 151 (1991)

Science http://www.sciencemag.org

ScienceWeek http://scienceweek.com