Stochasticity Versus Determinacy in Neurobiology: From Ion Channels to the Question of the "Free Will".

If one accepts that decisions are made by the brain and that neuronal mechanisms obey deterministic physical laws, it is hard to deny what some brain researchers postulate, such as "We do not do what we want, but we want what we do" and "We should stop talking about freedom. Our actions are determined by physical laws." This point of view has been substantially supported by spectacular neurophysiological experiments demonstrating action-related brain activity (readiness potentials, blood oxygen level-dependent signals) occurring up to several seconds before an individual becomes aware of his/her decision to perform the action. This report aims to counter the deterministic argument for the absence of free will by using experimental data, supplemented by computer simulations, to demonstrate that biological systems, specifically brain functions, are built on principle randomness, which is introduced already at the lowest level of neuronal information processing, the opening and closing of ion channels. Switching between open and closed states follows physiological laws but also makes use of randomness, which is apparently introduced by Brownian motion - principally unavoidable under all life-compatible conditions. Ion-channel stochasticity, manifested as noise, function is not smoothed out toward higher functional levels but can even be amplified by appropriate adjustment of the system's non-linearities. Examples shall be given to illustrate how stochasticity can propagate from ion channels to single neuron action potentials to neuronal network dynamics to the interactions between different brain nuclei up to the control of autonomic functions. It is proposed that this intrinsic stochasticity helps to keep the brain in a flexible state to explore diverse alternatives as a prerequisite of free decision-making.

A mathematical model of homeostatic regulation of sleep-wake cycles by hypocretin/orexin.

We introduce a physiology-based mathematical model of sleep-wake cycles, suggesting a novel mechanism of homeostatic regulation of sleep. In this model, the homeostatic process is determined by the neuropeptide hypocretin/ orexin, which is a cotransmitter of the lateral hypothalamus. Hypocretin/ orexin neurons are silent during sleep and active during wakefulness. Firing of these neurons is sustained by reciprocal excitatory synaptic connections with local glutamate interneurons. This feedback loop has been simulated with a minimal but physiologically plausible model. It includes 2 simplified Hodgkin-Huxley type neurons that are connected via glutamate synapses, one of which additionally contains hypocretin/orexin as the functionally relevant cotransmitter. During the active state (wakefulness), the synaptic efficacy of hypocretin/orexin declines as a result of the ongoing firing. It recovers during the silent (sleep) state. We demonstrate that these homeostatic changes can account for typical alterations of sleep-wake transitions, for example, introduced by napping, sleep deprivation, or alarm clock. In combination with a circadian input, the model mimics the transitions between silent and firing states in agreement with sleep-wake cycles. These simulation results support the concept of state-dependent alterations of hypocretin/orexin effects as an important homeostatic process in sleep-wake regulation, although additional mechanisms can be involved.

Propagation effects of current and conductance noise in a model neuron with subthreshold oscillations.

We have examined the effects of current and conductance noise in a single-neuron model which can generate a variety of physiologically important impulse patterns. Current noise enters the membrane equation directly while conductance noise is propagated through the activation variables. Additive Gaussian white noise which is implemented as conductance noise appears in the voltage equations as an additive and a multiplicative term. Moreover, the originally white noise is turned into colored noise. The noise correlation time is a function of the system's control parameters which may explain the different effects of current and conductance noise in different dynamic states. We have found the most significant, qualitative differences between different noise implementations in a pacemaker-like, tonic firing regime at the transition to chaotic burst discharges. This reflects a dynamic state of high physiological relevance.

Conductance versus current noise in a neuronal model for noisy subthreshold oscillations and related spike generation.

Biological systems are notoriously noisy. Noise, therefore, also plays an important role in many models of neural impulse generation. Noise is not only introduced for more realistic simulations but also to account for cooperative effects between noisy and nonlinear dynamics. Often, this is achieved by a simple noise term in the membrane equation (current noise). However, there are ongoing discussions whether such current noise is justified or whether rather conductance noise should be introduced because it is closer to the natural origin of noise. Therefore, we have compared the effects of current and conductance noise in a neuronal model for subthreshold oscillations and action potential generation. We did not see any significant differences in the model behavior with respect to voltage traces, tuning curves of interspike intervals, interval distributions or frequency responses when the noise strength is adjusted. These findings indicate that simple current noise can give reasonable results in neuronal simulations with regard to physiological relevant noise effects.

Stimulus-response curves of a neuronal model for noisy subthreshold oscillations and related spike generation.

We investigate the stimulus-dependent tuning properties of a noisy ionic conductance model for intrinsic subthreshold oscillations in membrane potential and associated spike generation. Upon depolarization by an applied current, the model exhibits subthreshold oscillatory activity with an occasional spike generation when oscillations reach the spike threshold. We consider how the amount of applied current, the noise intensity, variation of maximum conductance values, and scaling to different temperature ranges alter the responses of the model with respect to voltage traces, interspike intervals and their statistics, and the mean spike frequency curves. We demonstrate that subthreshold oscillatory neurons in the presence of noise can sensitively and also selectively be tuned by the stimulus-dependent variation of model parameters.

On episode sensitization in recurrent affective disorders: the role of noise.

Episode sensitization is postulated as a key mechanism underlying the long-term course of recurrent affective disorders. Functionally, episode sensitization represents positive feedback between a disease process and its disease episodes resulting in a transition from externally triggered to autonomous episode generation. Recently, we introduced computational approaches to elucidate the functional properties of sensitization. Specifically, we considered the dynamics of episode sensitization with a simple computational model. The present study extends this work by investigating how naturally occurring, internal or external, random influences ("noise") affect episode sensitization. Our simulations demonstrate that actions of noise differ qualitatively in dependence on both the model's activity state as well as the noise intensity. Thereby induction as well as suppression of sensitization can be observed. Most interestingly, externally triggered sensitization development can be minimized by tuning the noise to intermediate intensities. Our findings contribute to the conceptual understanding of the clinical kindling model for affective disorders and also indicate interesting roles for random fluctuations in kindling and sensitization at the neuronal level.

Fever induction by localized subcutaneous inflammation in guinea pigs: the role of cytokines and prostaglandins.

In guinea pigs, dose-dependent febrile responses can be induced by injection of a high (100 micro g/kg) or low (10 micro g/kg) dose of bacterial lipopolysaccharide (LPS) into artificial subcutaneously implanted Teflon chambers. In this fever model, LPS does not enter the systemic circulation from the site of localized tissue inflammation in considerable amounts but causes a local induction of the proinflammatory cytokines tumor necrosis factor (TNF) and interleukin-6 (IL-6), which can be measured in lavage fluid collected from the chamber area. Only in response to the high LPS dose, small traces of TNF are measurable in blood plasma. A moderate increase of circulating IL-6 occurs in response to administration of both LPS doses. To investigate the putative roles of TNF and prostaglandins in this fever model, a neutralizing TNF binding protein (TNF-bp) or a nonselective inhibitor of cyclooxygenases (diclofenac) was injected along with the high or low dose of LPS into the subcutaneous chamber. In control groups, both doses of LPS were administered into the chamber along with the respective vehicles for the applied drugs. The fever response to the high LPS dose remained unimpaired by treatment with TNF-bp despite an effective neutralization of bioactive TNF in the inflamed tissue area. In response to the low LPS dose, there was an accelerated defervescence under the influence of TNF-bp. Blockade of prostaglandin formation with diclofenac completely abolished fever in response to both LPS doses. In conclusion, prostaglandins seem to be essential components for the manifestation of fever in this model.