Monthly Archives: July 2014

THE EXCHANGE FROM A NETWORK PERSPECTIVE

The main purpose of a telecom network is to interconnect telephone users, devices and services. Since all users does not need to use the network all of the time, many network resources can be shared among several users.

A large exchange is a means to share switching re­sources by keeping many such resources in one common pool, thus providing statistical gains. A large exchange also will reduce network costs by enabling the sharing of net­work resources and reducing the number of lines, trans­mission capacity and external signaling needed if split into smaller exchange nodes or several specialized nodes.

The concept of an exchange can be related to the func­tions of a network node where many telecommunication lines are connected. However, the traditional functions as­sociated with an exchange are these days sometimes dis­tributed and located also to other node types in the net­work.

Functional and Structural Perspectives on Networks

A practical approach is to look at networks from the opera­tor’s perspective. From this point ofview a telecommunica­tions network can be divided into four major networks: an access network, a transport network, a signaling and con­trol network, and a management network. In more recently developed networks one often complements this view with service and media networks.

However, for different purposes and reasons, one may also talk about other types of networks that are related to each other in different ways. A reason to do this is that one often needs to look at networks both from functional and from structural perspectives.

In abstract functional perspectives, one views the net­work from the external environment as if it were a black box and focuses mainly on the services that the network provides. Functionally, a network can be defined as a set of points at which a set of functions are provided and at which certain properties of each such function can be measured. A network function at one level may be implemented by using a function or a set of functions at the adjacent lower level (see Fig. 1).

In structural perspectives, one takes a closer, more in­ternal, detailed and structural (white box) view and focuses on matters related to structurally and physically measur­able properties, including how the network functions can be partitioned, distributed in space, and allocated to net­work nodes. Structurally, one can hence define a network as a set of network nodes, each with a set of allocated node functions, placed and interconnected via logical intercon­nection points in this way enabling the node functions to cooperate so as to implement the network functions (or net­work services) to be provided.

Using the functional perspective of networks, one can distinguish some network concepts related to telecommu­nication exchanges. For example: an access network pro­vides access related functions at a set of network access points; a service network provides service (e. g., call) re­lated functions at a set of service points; a mobile network provides location and mobility related functions at a set of location points; a signaling network provides signaling functions at a set of signaling points; a connection net­work provides connection or switch related functions at a set of connection points; a management network provides management functions at a set of management points; a transport network provides transport functions at a set of transport points; and so on a media network provides in­formation functions at a set of media content provisioning points.

Figure 1. Network using functions from a lower-level net­work.

The functions at the network points mentioned above are functionally separated but can for economical or other reasons be colocated in the same physical node or site. The
networks are related to and use each other in many ways. For example: the signaling network is used to enable func­tional inter-work between the other networks; a subscriber can, by sending signals via the access network, obtain ac­cess to functions provided by the service network that es­tablishes these services using (by sending of signals to) the connection network(s); the transport network is used by all the other networks as a bearer for transport of raw data.

ENDOCRINOLOGY

Cells communicate by means of electrical coupling (e. g., direct contact interaction or neurocontrollers) and chem­ical interaction (i. e., endocrine control often for the long­distance type). These communication networks regu­late vital processes such as growth, differentiation, and metabolism of tissues and organs in all living systems. The nervous system and the endocrine system are linked by the integrating function of the hypothalamus, a specific region of the brain.

The specific endocrine contribution to regulation forms the subject of this section. It involves signal genera­tion, propagation, recognition, signal transduction, and re­sponse to a particular stimulus. In endocrine signaling, the pertinent cells release substances called hormones. The main endocrine glands are the hypothalamus, the anterior and posterior pituitary, the thyroid, parathyroid, adrenal cortex and medulla, pancreas, and the gonads (i. e., ovaries or testes). Hormones can be chemically classified into three groups: steroid hormones (e. g., estrogen and progesterone), peptide or protein hormones (such as insulin, prolactin, and the various releasing hormones), and a category de­rived from the amino acid tyrosine (i. e., thyroxine and tri – idothyronine). These substances are usually transported via the bloodstream. Therefore they are distributed within the whole body. Lifetime in plasma may range from sec­onds to days. In order to ensure a selective action at a special site, target cells exhibit sensitivity for a particu­lar hormone and only in these cells will an appropriate response be induced. The ability of target cells to exclu­sively respond to specific hormones depends on the pres­ence on the membrane of receptor proteins that are unique for a particular hormone. The action may take place on the cell surface (with or without an intracellular mediator— the second messengers, substances that serve as the relay from the plasma membrane to the biochemical machinery inside the cell), or only within the interior of the cell, at the nucleus level.

Besides negative and positive feedback circuits, the en­docrine system features rhythmic control by pulsatile re­lease of certain hormones leading to circadian variations. Also, endocrine control can be pharmacologically modified with the use of synthetic hormones. The modern field of molecular biology is also concerned in part with many aspects of the physiology of the endocrine system at the (sub)cellular level TELECOMMUNICATION EXCHANGES

The telecommunications network including both public and private parts is one of the most important systems created by humans in modern civilization. The network enables people to communicate between continents at all times of the year and allows thoughts and ideas to be ex­changed between families, companies, and governments. Several phone calls have changed history. The network also saves lives every day. An example is when mobile sub­scribers in their cars call emergency numbers when wit­nessing car accidents.

Due to the network’s great importance the requirements on the components that constitute the system are several. The more central parts of the network must always be available. The most important type of such a central com­ponent in the classic telecommunication network is the telecommunication exchange. The exchange enables many calls to be switched and alternative paths to be taken if a path in the network has failed, and hence should be very reliable. This reliability is probably the most important re­quirement of the exchange. Other requirements are that the exchange should be able to cope with all signaling stan­dards in a network. It should also be able to coexist with all equipment in the network, irrespective of its age or fabri­cation. This imposes great requirements on backward com­patibility.

Note that a telecommunication exchange is sometimes also called a central office or switch. There are both public exchanges and private branch exchanges. An exchange is a node in a telecommunications network that includes func­tions for access, control, switching and charging of calls. These parts may be physically separated and distributed, also to places outside the central office site; and act as subnodes connected via signaling protocols, for example be­tween access, control and switch resources. A call in a mod­ern context is not just about phone calls but is a general connectivity service for two or more parties that need ac­cess to communication bandwidth over a shorter or longer period of time. Connection services can be established not only between peer users but also from peer users to central servers.

A telecommunications exchange is one of the most com­plex systems created by humans. State-of-the-art hard­ware and software technology is used to create very large systems in terms of the number of connected subscribers and number of switched calls. The development cost of an exchange is very large. Several millions of hours of hard­ware and software development are invested each year. An exchange can connect hundreds ofthousands ofsubscribers and switch a million of calls at the busiest hour during the day. And all of this is in real time, meaning short setup time and low delay of the voice. The quality of service require­ments is thus great. Whether or not the call is made be­tween two different continents, the call has to be switched through with practically no delays and provide an accept­able speech quality.

The basic service associated with traditional telephony is the enabling of the bidirectional voice communication di­alogue between two persons located at different but fixed places. To enable such voice communication each person uses a telephone set equipped with a transmitting device starting with (1) a microphone that translates acoustic en­ergy to electric energy and (2) a receiving device ending with a loudspeaker that does the reverse transformation.

Each person who has access to telephony also wants to be able to select whom to connect to and have a dialogue with. For the establishment of such a connection service, one must be able to signal from each place to any other se­lected place that a connection is wanted between one call­ing and another called person. The traditional call is hence actually directed to a place where a called person is sup­posed to be located. To enable the sending and reception of alert signals the telephone set is equipped with a device that can generate some form of electrically encoded signal and a device for reception and transformation of such a signal to an acoustic signal. For a basic two party dialogue, two telephone sets are connected, usually electrically via a pair of wires. In the very beginning the second wire was im­plemented by an earth connection; this, however, resulted in a high level of cross-talk when several such unshielded wires came close to each other.

To start with, telephones were mainly used between two or a few locations such as between a shop owner’s office, workshop, and home. A simple n-way switch could be used to select who to talk to within such a small mesh structured private net. Later when several families and organizations in a town had such connections, the communication pos­sibilities were extended. This was done by connecting all wires in a star structure to a central office equipped with an exchange consisting of a manual switchboard and an operator. Hence, each telephone in such a local area was connected via a single wire or a pair of wires to the switch­board. Such a pair of wires is often called a subscriber (or access) line or loop. The switchboards themselves were then interconnected via other pairs of wires, called trunks. To enable voice signals to be carried with less distortion and over longer distances (by use of inductor coils and am­plifiers respectively) the trunk lines soon came to use four wires, one twisted pair in each direction.

The rapid growth of the telephony network during the early days may well be explained by the fact that the net­work technology had large similarities with the technol­ogy already used for telegraphy. But perhaps most im­portant, the telephone set replaced the need for a tele­graph operator knowing Morse code and thus simplified the human-machine interface. To establish a call, the call­ing person had first to send a ring signal to the operator and then tell the thus alerted operator whom to be connected to. If the call was local, the operator then could send a ring signal to the person being called and ask this person for permission to set up a connection to the person calling. The operator could at this point make clear how or by whom the call was to be paid. A call could then be established by con­necting the two pairs of wires to each other. To simplify the operator’s work the manual switchboard was designed to make it simpler for the operator to supervise the lines that were busy and to handle both the request for, establish­ment and the ending of a call. If the call was non-local, the operator first searched for and selected a free trunk in the right direction and then called one of the operators at the receiving end of the trunk (using it as a signaling trunk) and asked for help to establish the call. Such searching, routing and forwarding tasks presume that that the oper­ator had some knowledge about the networks topology. For a long-distance call a chain of operators had to be involved via a chain of trunk line links before the called person could be reached and asked for permission to set up the call via voice trunks along the same path.

To set up a long-distance call, a number of resources must be free. Those resources mainly were operators and voice trunks. If some resource was lacking, the operator could serve the customer by organizing a waiting list or job queue and then set up calls when resources eventually became free. The operators had to keep track of each job by help of a written job record including time stamps for charging purposes and also provide time supervision of the ongoing calls to guarantee that resources allocated to a call were released even if the calling parties forgot to send an end-of-call alert signal.

When the number of subscribers increased, quite a large number of operators could work in the same ex­change office. Different techniques were then used to en­sure that they could share the workload and coordinate the setup of calls between subscribers connected to different switchboards. Parallel processing of call attempts could be achieved, for example, by distributing the incoming access attempt from a subscriber to a non-busy switchboard. An­other example is that one single operator could handle the setup of a local call within a large exchange consisting of many switchboards by use of multiple point technique — that is, by having switchboards all outgoing lines connected to all other switchboards in the office.

When the demand for telephony increased, the num­ber of operators swelled. Pressure grew to decrease the costs for human switchboard operators but also the time to set up calls. Requirements on increased personal integrity were also a reason to try to automate the call setup pro­cedure. The first automated switches were based on the use of electromagnetic coils, effectuating drive mechanisms and contact points. These electromagnetic devices were, in turn, controlled by signals generated from the telephone set by use of a dial. The dial could generate sequences of current pulses, where the number of pulses corresponded to a dialed decimal digit. A subscriber was given a tele­phone number related to a corresponding access line. The digits used to represent this number controlled the behav­ior of the switch. The signals were first decoded and used directly to control the switch movements in a decadic way and somewhat later indirectly via registers. The use of reg­istered signals (1) reduced the requirements on timing of the signals, mechanical precision, and preventive mainte­nance and (2) increased the flexibility by making number translations possible.

All decisions to be made were not controlled by digits. For example, a number of trunk lines between two ex­changes could be treated as a group and the digits used merely to select and seize a trunk group going in the right direction. Then one could do a search (or hunt) for a free trunk line and, when available, select one within this group. When no trunk line was free, another possible route could be tried. However, the number of alternative routes was limited both due to economic reasons and be­cause the networks soon came to be built more or less hi­erarchically to simplify the coordination work needed to establish non-local calls. If no free route could be found, a blocking situation occurred and the call attempt had to wait for resources to become free. The operator could then use different methods to supervise resource release events and to handle the queue of waiting call attempts.

It is interesting to note that the tasks performed by a human operator, such as searching for, reservation, moni­toring and management of resources and charging records, were very similar to what the control system of a mod­ern exchange does. There are also interesting analogies between the call routing, redirection and answering ser­vices provided by a human operator and what today can be provided by the control system of an automated exchange.

The network has evolved from very simple bidirectional communication links via small private mesh and star net­works that as soon as signal regeneration and amplifica­tion technology permitted were interconnected via transit networks to more public, global and hierarchical network structures. The number ofindividual trunk lines could also later be reduced by multiplexing techniques allowing sev­eral calls to share a physical line. This evolution started with frequency division and has evolved via digital time division toward many different combinations of frequency, phase, time and code division. Multiplexing creates a logi­cal network layer on top of the physical transmission media implementing a number of logical lines (or channels) and hence a more efficient use of each physical line.

Logically, signaling has always been separated from the voice connection. A trend has been to clarify this by a sep­aration into a signaling trunk network and a voice trunk network. However, the signaling network may in practice use reserved logical lines or channels multiplexed on top of the same physical lines as the voice network.

Voice encoding, with its influence on transmission and switching, has evolved from analog; via digital to com­pressed digital representation and the signal encoding has evolved from simple current pulses, via frequency-encoded signals to digital message records.

Manual exchanges handled by human operators were quite flexible and intelligent in many ways since the prim­itive alert signals simply could be complemented by verbal communication between subscriber and operator — that is, human to human. Automation required a predefined sig­naling scheme, including not only simple alert signals but also encoding of the phone number of the called line. To make the automated exchanges able to provide large flex­ibility and more advanced services, the first decadic con­trol of an exchange directly from signals representing dig­its evolved via register mapped control to stored program control of exchange behavior. The automation possibilities were increased further by introduction of larger signal al­phabets and protocols capable of more than just alert sig­nals and digits.

A great step from a functional point of view was the in­troduction of radio transmission for cellular coverage and wireless access to/from mobile terminals. This allows a call to be directed to a mobile terminal carried by a person rather than just to a fixed place (terminating a wire or fiber). Technically it was not new to reuse the radio spec­trum by dividing a geographical area into regions, in this case called cells, but connecting the cells covered by base stations to the switched telecommunication network was new and created many new challenges and opportunities.

Other steps in this direction are: digital subscriber lines; new call services (often for redirection of calls) sometimes substituting what a manual operator previously could give help with; introduction of personal numbers that in con­junction with mobility services can help to make it easier to reach a specific person rather than a phone terminal and also can increase competition among operators if the per­sonal phone number becomes a property of a person rather than an operator; the merger of telecommunication and data communication networks that enables new multime­dia communication services.

A recent such network convergence that will simplify the evolution of streamed and real-time multimedia ser­vices is the sharing of a more and more common infrastruc­ture for fixed and mobile voice, data and video services. In this new setting some of the functions of a telecommunica­tion exchange become obsolete and other more dedicated nodes become more important. For example, circuit switch functions may be replaced by packet forwarding while new forms of access control, authorization, authentication, ad­dress location and charging services will be more impor­tant. This development along with international standard­ization is also believed to increase the competition between different equipment suppliers.

An exchange can include support for almost all func­tions in a telephone network. However, one can also dis­tinguish specialized network nodes. Early examples of these specialized nodes are local exchanges and transit exchanges. Today one can also distinguish other types of nodes such as network access nodes; switch, call and ser­vice control nodes; mobile switching nodes; and network database nodes (for example, databases for: number trans­lations; handling of subscriber service profiles; location in­formation supporting mobility of subscribers, authentica­tion/authorization data, equipment data). Other nodes are different types of information service, media content, e – trade transaction and access right handling servers.

FRACTAL AND WAVELET ANALYSIS OF CARDIOVASCULAR PARAMETERS

The design and function of the cardiovascular system is commonly described in terms of a muscular pump and a network of branching vessels. This traditionally implies an interpretation in terms of the laws of physics (12,13). More recently, analysis using the fractal properties of physiolog­ical observations has been employed. This appears to have a great potential for the study of physiology at scales of resolution ranging from the microcirculation to the entire organism. Other new techniques include wavelet analysis and nonlinear approaches. They are mentioned here to let the reader become aware of the existence of these tools.

Fractal analysis

The values of the measured properties of many physiolog­ical systems seem random. This view originates from the tools that have been used when analyzing the details of or­gans. Nowadays there are new methods to analyze seem­ingly random experimental data, such as a time-series ap­proach. These methods use many of the properties and ideas of fractal analysis. The mathematician Mandelbrot chose the word fractal to denote objects or processes with multiple-scale properties, that is, the ever finer subdivi­sions of objects or processes as they are viewed at progres­sively higher magnification. Until recently, our ability to understand many physiological systems was hampered by our failure to appreciate their fractal properties and to in­terpret scale-free structures (29).

Wavelet analysis

Fourier analysis (i. e., decomposition of a signal into sinu­soidal waveforms) is not suitable for signals with discon­tinuities. This type of computational headache can be suc­cessfully treated by wavelet (ondolettes) analysis, a pow­erful technique developed by Mallat and Meyer in France (30). Later, Daubechies (at Princeton) discovered the dual family, representing the high-frequency range and the smooth parts (low frequencies). Wavelet analysis owes its efficiency to the fast pyramid algorithm, reducing the num­ber of calculations by down-sampling operations, that is, steps that remove every other sample at each operation (halving the data each time). The method has vital appli­cations in compression ofsignals and images, in addition to noise reduction (a common problem in biomedical record­ing, in particular in magnetic resonance imaging). The cur­rent popularity of wavelet analysis may be exemplified by the commercially available software “Wavelets for Kids” developed at Duke University.

Nonlinear analysis

Heart rate (HR) and blood pressure (BP) are controlled by several central nervous system oscillators and differ­ent control loops (see Fig. 7). Interactions among these units may induce irregular time courses in the processes they govern, but the underlying subprocesses also include deterministic behavior. These irregular time courses can be more accurately characterized by dynamic nonlinear analysis rather than by linear time series. Typically, one­dimensional time series data are transformed into multidi­mensional phase-space plots, thus filling selected regions. Two essential aspects of such plots include:

1. The correlation dimension, which is a measure of the complexity of the process studied, that is, the distri­bution of points in the phase space

2. The Lyapunov exponent, a measure of the pre­dictability of the process, quantifying the exponential divergence of initially closed state-space trajectories.

CONTROL OF HEART AND CIRCULATION

The tracing of an electrocardiogram (ECG) is one of the most popular icons displayed whenever referring to the field of medicine. From a clinical point of view, the mea­surement of the ECG is attractive because it involves a noninvasive procedure that can be performed quickly using either portable or fixed equipment. The cardiac system can best be described as a mechanical pump system that is trig­gered and synchromized by electrical signals. Because of its noninvasive nature, the ECG is helpful in providing some preliminary information on cardiac rhythm, electrical con­duction, and its disturbances. Figure 6 shows the preferen­tial pathways for conduction of the excitation wave, which closely follows functional anatomy of the heart, that is, starting at the sinus node in the right atrium, traveling towards the apex of the ventricles, followed by a spread to­wards the outflow tract. This route implies that the blood

Figure 6. A cross-sectional view of the heart, showing the two ventricles and both atria, along with the conduction system for the electrical impulse running from the sinus node towards the apex and then upwards along the muscular walls of the ventricles.

is expelled from the ventricles similar to the optimal direc­tion for squeezing out the contents of a tube of toothpaste, that is, running from the very bottom up towards the open­ing at the opposite site.

The rhythm of the healthy heart is not constant but ex­hibits a certain degree of HRV (20). Many studies have em­ployed advanced mathematical techniques to analyze the cardiac rhythm and estimate the relative contribution of the sympathetic and parasympathetic drives, respectively (21, 22). But the perpetual change of the rhythm has also profound mechanical consequences: An increase in cycle length implies facilitation of ejection both by increased fill­ing (i. e., elevated preload) and reduced opposing pressure at the time of the valve opening (i. e., lower afterload), while during the next beat with a shorter interval the opposite applies. In other words, impeded and facilitated beats ap­pear to alternate, thus possibly improving stability of the complete circulatory system.

The left ventricle has often been modeled as a sphere or prolate ellipsoid, but neither geometry conforms with reality. Independent of geometrical assumptions, Beringer and Kerkhof (23) have shown that a fairly linear relation­ship exists between end-systolic volume (ESV) and end – diastolic volume (EDV). This notion has been verified in human patients and also in their experimental investiga­tions when studying the volume regulation in physiolog­ically operating chronically instrumented dogs (24). The regression coefficients appear to be characteristic of ven­tricular volume regulation and are sensitive to inotropic intervention (i. e., adrenergic agonists and blockers). This relationship implies that a clinically important cardiac per­formance indicator, namely ejection fraction (EF) is in­versely related to ESV (25). Furthermore, using the Suga – model, myocardial oxygen consumption can potentially be predicted from a single noninvasive determination of ESV Quite similar to the description of pulmonary dynamics, the dynamics ofheart and vessels are also characterized by pressure-volume (P-V) relationships. Flow ( Q) equals the time derivative of a changing volume. Another frequently employed derived index is elastance ( E), the reciprocal of compliance, which is defined as the material property that enables resisting deformation. In analytical form,

where V0 is the unstressed volume. This is analogous to Hooke’s law (length-tension relationship), which states that tension (i. e., force per unit area) equals Young’s mod­ulus times the relative change in length.

Spontaneous oscillations of arteries have also been ob­served, with a period ranging from 45 s to 60 s for the ra­dial artery diameter under resting conditions, resulting in an almost twofold change in distensibility (26). A general model of the properties of a muscular blood vessel at vary­ing levels of contraction predicts that the vessel becomes unstable at high levels of contraction (27). A lumped model of the arterial tree can also be considered, similar to the lumped model for the lungs as presented before. Imagine that the arterial bed with all its branches is replaced by a single vessel with a diameter similar to the human aorta and encompassing an identical hemodynamic resistance ( R). How long would this hypothetical noncollapsible aorta be? From Poiseuille’s law we know that R = 8nl/nr4. If the radius r =1 cm, the blood viscosity n = 3 cP, and R is the mean driving pressure divided by mean flow, that is, 100/5 mmHg-min/L, then the length l turns out to be approxi­mately 200 m.

Obviously, the real circulation consists of various vas­cular beds that supply blood to the various organs such as brains, kidneys, skeletal muscles, and intestines. One pow­erful regulatory mechanism used to accommodate varying needs in a single organ is based on the principle of redis­tribution of flow within the body by local vasodilation and constriction of arterioles. Another phenomenon, called au­toregulation, is based on vasoactivity that maintains the flow at a particular level even during acute changes in pres­sure.

Arterial blood pressure is well controlled by several sys­tems. Figure 7 presents a survey with emphasis on time course and gain for each subsystem involved (28). Hyper­tension can then be interpreted as a disorder in which the set point is altered while the control systems continue to regulate towards an erroneously determined set point. This view also explains the chronic nature of a condition in which blood pressure even at rest is elevated. Going back to Fig. 1 this abnormality would be equivalent to a situation in which the heater feedback system continues to maintain a given (wrong) set point.

CLOSED LOOP DRUG DELIVERY

The field of pharmacokinetics concerns the analysis of fac­tors that affect absorption, distribution, and elimination of drugs in the body. In contrast, pharmacodynamics refers to interactions at the receptor site. Drugs may primarily act locally (the so-called topical type such as eye drops or inhaled aerosols) or they may more or less simultaneously (intentionally or not) influence many organs in the whole body (i. e., systemic type). Obviously, temporal as well as spatial considerations are relevant in pharmacokinetics. The route of administration mainly determines where a pharmacological substance exerts its primary action. Be­cause of the frequently occurring side effects of almost all drugs, it is also important to gather information about the distribution and uptake of a drug at a site that is not the primary target. Apart from the site of administration, it is important to have some estimate about the optimal speed of delivery at the desired location. Injection into the blood­stream is fast and potentially transient, while application on the skin (transdermal route) is slow but has the advan­tage ofbeing long-acting and rather constant. In particular, in iontophoresis a drug is delivered transdermally by using an electric field to enhance the transport of small, poorly absorbed ionic drugs across the skin surface with the ad­vantage that only a low dosage of the drug is required (16). All these considerations make clear that it is useful to de­velop models that incorporate the various (anatomical and physiological) compartments that are spatial and chemical in nature in order to predict the concentration and time pattern at a target site dependent on the location of in­troduction of the substance, as well as the particular time sequence (e. g., bolus versus repeated doses) of administra­tion.

VISION RESEARCH AND EYE MOVEMENT

Various aspects of eye movement have been extensively described in PHYSIOLOGICAL MODELS, DEVELOPMENT by T. Bahill. Therefore, we focus here on our main theme, namely fluctuations, instability, and variability as observed in the oculomotor control system. Basic concepts of the neuro­muscular system controlling movement of hands and limbs also apply to eye tracking. The task of permanently center­ing a moving image on the fovea is realized by a multilevel system involving reflex loops, volitional control, and pre­dictors. Saccades are rapid successions of conjugate steps of eye rotation that permit the positioning of a target image onto the fovea. The eyes voluntarily move from one fixation point to another, as can be observed during reading. They are preceded by a reaction time of about 0.20 s and follow a typical course of rapid acceleration and subsequent de­celeration with occasionally a small overshoot. Accommo­dation (focusing) is driven by a blur of the target image on the retina. Smooth pursuit involves a slow but continuous following movement needed to perform a smooth tracking task. Disjunctive movements of both eyes permit vergence (binocular fixation system resulting in convergent or diver­gent movement). Accommodation and vergence form an in­teractive dual-feedback system. Physiological nystagmus (min eye movements) are repetitive fast and slow move­ments that adds to the visual acuity by preventive bleach­ing while shifting to different receptors. Acceleration ofthe body requires the vestibulo-ocular reflex to become oper­ational. It is believed that (white) noise enhances stabil­ity (as in HRV), while fluctuations affect accommodation. Hung described a nonlinear static model, containing the depth of field (as a dead-space operator for accommoda­tion) as well as Panum’s (17) fusional area (as a dead-space operator for vergence), and found that these operators are able to account for the discrepancy between results using the phoria and fixation disparity methods (18).

RESPIRATORY CONTROL

During inspiration, air enters the lungs from the nasal pas­sages (conchae) or the mouth via a branching system of tubes ending in numerous small but highly elastic hollow structures (alveoli). These elastic elements are in contact with small blood vessels and are therefore the sites of gas

Figure 5. There is increased local obstruction at point C in case of asthma, whereas in emphysema the trajectory between C and the alveoli shows a permanent decrease of diameter.

exchange. Expiration implies transport in the opposite di­rection, from the lungs towards the final tube (the wind­pipe or trachea). Expansion of the lungs is normally real­ized by muscular activity of both the diaphragm and the intercostal muscles (19).

The dynamics of respiration are commonly described in terms of a pressure-volume relationship (to study restric­tive diseases such as interstitial fibrosis and pulmonary edema) and derived quantities such as flow ( Q, to study obstructive diseases such as lung emphysema) and com­pliance (the ratio of volume changes resulting from varia­tions in pressure). In contrast to emphysema, asthma is a reversible obstructive airway disease, because it is caused by an increase in smooth muscle tone in the large bronchi. Figure 5 illustrates the nature of these abnormalities in a lumped parameter model, consisting of the thoracic wall ( T), an overall spherical elastic element with alveolar pres­sure ( Palv) inside, atmospheric pressure ( Patm), pleural pressure ( Ppl), total airway resistance ( Raw), and C is the usual point of collapse of the airways acting as a Starling resistor (i. e., a collapsible tube affected by the pressure of its surroundings). It can be derived that

Q (inspiration) = (Pstm — Palv >/tfSw = —Q(expiration)

The stimulus to breathing is controlled by a particular area in the brain stem called the respiratory center. The rhythmical contraction of the diaphragm and intercostal muscles determine the inspiration sequences, that is, the brain generates alternating cycles of firing and quiescence in the responsible motor neurons.