ALBERT

Notes: Summary The subject of this study has been the way in which fast writing movements are guided. It is concluded that: 1. For fast handwriting the principle of position feedback does not hold; 2. There is physiological evidence for resolving the writing movements into two more or less perpendicular directions; 3. The shape of a word is determined by the timing of the muscle contractions and not by the magnitude of the forces used; 4. A general change of size is coupled to a proportional change in the magnitude of the forces.

Notes: Summary The performance of the Learning Matrix (LM) is suitable for the design of adaptive networks of higher complexity. It has been published, how to connect a LM with a generator of patterns (binary or nonbinary) and a ring-counter to result in an automatic classification of the presented patterns. This paper describes, how to connect two LM's to form an “Autonomous Learning Matrix Dipole” (ALD) and how to organize it, so that it adapts itself to an environment according to a given evaluation scale. For this purpose, a third type of input (beside “e” and “b”), namely “h” seems to be useful. This h-input controls the rate of adaptation of the LM. Using such ALD's, one may design adaptive structures of even higher complexity, for example with an adaptive internal model. The principle of “Learning Matrices” has been explained in detail (see e.g. IEEE Transactions on Electronic Computers, Vol. EC-12, No. 6, December, 1963, pp. 846–862). Using such “learning matrices” (LM), one may build up adaptive networks with rather interesting functions. Perhaps they are interesting for the physiologist and psychologist as well as for the engineer. Let us first recall the most essential details of the LM's.

Notes: Abstract Small-signal frequency-response measurements of a locust visual reflex in the range 0.0014 cps to 6 cps suggest description via special departures from ordinary linear-system characterization, at both low and high frequencies. At low frequency the response is similar to that of an adapting system having the formal characteristics of fractional-order differentiation. Adjustment of test-pattern luminance from 0.00025 to 1 lambert increases the order of this differentiation by about 0.37. Comparison of intensity-dependent properties of the reflex with related electrophysiological studies of photoreception implies that the frequency dependence of the optomotor response is dominated by receptor dynamics. At high frequency, phase lag is less than that required by minimum-phase linear description of the gain curves. In order to avoid the question of “predictive tracking” by the locust, a time-and frequency-dependent gain modulation is suggested to account for the less-than-minimum phase property, and shown to be qualitatively compatible with transients in gain at the start of sinusoidal pattern motion. The motion-perception model of Hassenstein and Reichardt, shown previously to predict the gross gain and phase relations of the small-signal locust response, is apparently an inappropriate description of these data since the decisive gain and phase measurements are more plausibly accounted for by intensity-dependent adaptation than by the properties of the model underlying the original prediction. An abstraction of the fundamental properties of their multiplicative interaction of ommatidia, however, can assist electrophysiological search for neural correlates of motion perception; it is shown that, under certain conditions, interaction of receptor channels via lateral shunting inhibition is (a) indistinguishable from the multiplicative motion-perception topology, and (b) formally related to a nonlinear property of inhibition in the eye of Limulus.

Notes: Summary Pattern-specific response in the visual system is represented in terms of a mathematical model and as the response of corresponding sequential circuits. The mathematical model employs standard relations of propositional calculus, with extension to the time domain. It is shown that pattern sensitive units of many types can be generated by successive applications of a generalised contrast operator together with a spatial summation operator. The implications of the models are interpreted in neural terms.

Notes: Summary The maintained activities of 116 neurons of the cat's optic nerve were recorded on tape. Non-sequential interval histograms were computed; in 29 neurons a multimodal type was observed. The peaks were regularly spaced corresponding to a fundamental period of 20–35 msec. Similar interval histograms were found also during constant illumination and under the influence of strychnine and picrotoxin. In the same neuron the interval histogram may alternate between a unimodal and a multimodal type. The discussion is based on the assumption of two interacting generators differing with respect to their characteristics.

Notes: Summary In this paper the effect of the statistical properties of the ganglion cell discharge on the transfer characteristics of the cat's retinal sensory system is studied. On the basis of results reported by the literature, a mathematical model of the system is defined. The model is then studied by digital computer, to obtain the amplitude of its response to sinusoidal stimulation as a function of frequency. The results show that, as the discharge is not Poisson-like, a positive resonance exists between stimulus and discharge at stimulus frequencies whose period is of the same order as, or smaller than the mean interval of the discharge. The amplitude of the resonance is a function of the statistical parameters of the discharge. These results fit well experiments recently carried out.

Notes: Summary Self-improving control systems may belong to either of two categories, according to whether or not they embody an explicit model of the part of their environment with which they interact. The two forms of operation are discussed and compared, and it is shown that the two may be mathematically equivalent. The treatment also gives theoretical justification for a particular mode of operation for nonmodel-forming controllers.

Notes: Zusammenfassung Man weiß heute noch wenig über das Verhalten dynamischer Systeme, die aus vielen gegenseitig aufeinander einwirkenden Komponenten bestehen, und über die Faktoren, die ihr Verhalten im ganzen und nicht so sehr in Einzelheiten bestimmen. Diese Arbeit beschreibt eine Untersuchung solcher Systeme, die aus Einzelelementen aufgebaut sind, welche rekursive logische Funktionen berechnen. Jedes Element hat zwei binäre Eingänge und einen binären internen Zustand, der gleichzeitig den Ausgangszustand des Elements darstellt. Der Ausgang des Elements kann vervielfacht werden. Rekursion wird eingeführt, indem man den Zustand eines Elements im unmittelbar folgenden Moment der Systemzeit (t+1) abhängig macht sowohl von den beiden momentanen Eingangszuständen als auch von seinem momentanen internen Zustand zur Systemzeit (t). Daraus folgt, daß 256 verschiedene Funktionen berechnet werden können; ein Element wird dadurch bestimmt, daß es genau eine dieser Funktionen berechnet. 100 identische Elemente, von denen jedes dieselbe Funkion berechnet, stellen ein System dar. 256 verschiedenet Systemtypen, die mit den 256 logischen Funktionen übereinstimmen, werden mit Hilfe einer digitalen Rechenmaschine studiert, indem man 5 verschiedene Verbindungsarten benützt und das System in 10 beliebig gewählten Anfangszuständen beginnen läßt. Nachdem der Angangszustand einmal hergestellt ist, zeigt das System, ohne weiteren Eingriff von außen, das ihm eigene Verhalten. Wir haben uns besonders damit beschäftigt, die Einwirkung der Faktoren auf das Verhalten der Systeme zu studieren, die bestimmend sein könnten für (i) die Zeit, die ein System braucht, um seinen terminalen Cyclus zu erreichen, und (ii) die Länge der Periodizität des terminalen Cyclus. Unter den Ergebnissen der Untersuchung fallen die folgenden besonders auf: 1. Diese Systeme zeigen eine Tendenz, in einem komplizierten Verhaltungscyclus zu terminieren. Der sehr kurze terminale Cyclus ist keineswegs die Regel. 2. Der Verhaltenstyp — von Einzelheiten abgesehen — ist oft überraschend unabhängig von den Verbindungen. 3. Ein entscheidender Faktor für die Länge der Zeit, bevor ein Terminalcyclus festgestellt werden kann, ist das Ausmaß, in dem die Elemente informationsverarbeitend wirken. 4. Ein Faktor, der die Tendenz zu einem sehr kurzen Terminalcyclus maßgeblich beeinflußt, ist die Anzahl der Bedingungen, unter denen die Zustände der Elemente bei aufeinanderfolgenden Zeitintervallen unverändert bleiben. 5. Elemente, deren Übergänge von ihrem vorhergehenden Zustand abhängig sind, begünstigen kurze Einschwingperioden vor langen Terminalcyclen. Die Bedeutung dieser Ergebnisse in ihrer Anwendung auf biologische informationsverarbeitende Systeme wird besprochen.

Notes: Zusammenfassung Zweck des vorliegenden Aufsatzes ist es, die Form des Spektrums von Sprechsignalen, wie z.B. von Selbstlauten, zu erklären.

Notes: Summary A simple neuron model has been developed by adding simple linear excitatory and inhibitory synapse models to Hodgkin and Huxley's model of the giant axon. Some cases of pulse processing of this model have been investigated by an electronic digital computer. The postsynaptic response of the model depends on the shape and on the intensity of the presynaptic impulses. Therefore, it has been activated by single and double presynaptic impulses, by pulse trains and by DC excitation (inhibition). DC excitation (inhibition) simulates the case that there are very many excitatory (inhibitory) synapses activated alternately by pulse trains of high repetition rate. The characteristic curves of the model for DC excitation (inhibition) represent the relationship between the presynaptic DC excitation (inhibition) and the depolarisation (hyperpolarisation) of the postsynaptic membrane potential (Fig. 5). These curves show saturation. At the beginning they are rather steep. With increasing depolarisation (hyperpolarization) the steepness decreases. With DC excitation four states of different stability can be distinguished. Up to 4 mV depolarisation of the postsynaptic membrane the model is absolutely stable; no actionpotentials can be evoked. A quasi stable state exists for depolarisation between 4 mV and 8.8 mV. In this range single actionpotentials or trains of actionpotentials may be evoked. For depolarisations between 8.8 and 18 mV the model is instable; each displacement of the membrane potential in this range generates a periodic tram of impulses. Another quasi stable state exists for depolarisations above 18 mV. In this range single actionpotentials may be evoked which are followed by a damped oscillation (Figs. 21, 23). The characteristic curve for inhibition by single presynaptic impulses represents the relationship between the amplitude of the IPSP and the intensity of inhibition (in the neuron: number of synchronously activated synapses; in the model: amplitude of presynaptic impulse). Inhibitory impulses have been superimposed on DC excitation or DC inhibition. The amplitude of the IPSP increases with increasing depolarisation and decreases with increasing hyperpolarisation (Figs. 12, 14). The characteristic curve for excitation by single presynaptic impulses gives the relationship between the amplitudes of the postsynaptic response and of the presynaptic impulse (Figs. 7, 27). This curve raises slowly in the subthreshold range, very fast close to the threshold and remains almost constant above threshold (Fig. 27). If subthreshold presynaptic impulses are superimposed on DC excitation or inhibition, the amplitude of the postsynaptic response (EPSP) decreases in both cases (Figs. 9, 10). Because of the nonlinearity of the model there are many eases where postsynaptic responses of single presynaptic impulses cannot been added linearly (Figs. 15–20). In general actionpotentials are evoked by sufficiently high excitatory input impulses. They also may be evoked by sufficiently high inhibitory input impulses. In this case the first positive afterpotential may exceed the threshold and evoke a spike (Fig. 30). The model has been activated by periodic pulse trams. Depending on the input intensities and on the repetition rates the model shows three modes of operation: 1. full (one to one) transmission, 2. partial transmission (division of repetition rate), 3. no transmission at all (subthreshold activation) (Figs. 32, 33).

Notes: Summary In isolated receptors the impulse frequency following “step” stretches had a highly significant correlation with both muscle length and tension; any deviations from linearity were in opposite directions, impulse frequency rising more quickly than linearly with length and more slowly than linearly with tension. The impulse frequency decayed according to a power function of time from application of a step increase in length. A transfer function was derived and used to predict responses to sinusoidal and constant velocity stretches. The experimental data generally agreed with predictions. The deviations that were found could be accounted for by considering quantitatively any non-linearity between frequency and length, the adaptation of the impulse frequency to constant currents, the all-or-none nature of the action potential, and the viscous forces present during dynamic stretch. The approximately linear relationship between impulse frequency and muscle length and muscle tension is discussed. Muscle tension appears to be the more direct causal agent of impulse generation. Possible physical bases for the transfer function are also considered.

Notes: Summary Recent electrophysiological data obtained from auditory nerve fibers of cats have made possible the formulation of a model of the peripheral auditory system that relates the all-or-none activity of these fibers to acoustic stimulation. The components of the model are intended to represent the major functional components of the peripheral system. These components are: (i) a linear mechanical system intended to represent the outer, middle, and mechanical parts of the inner ear; (ii) a transducer intended to represent the action of the sensory cells; and (iii) a model neuron whose properties are intended to represent the nerve excitation process. A general-purpose digital computer has been used to determine the response of the model to a variety of acoustic stimuli. These results have been compared with data obtained from auditory nerve fibers.

Notes: Summary The human pupillary response in one eye can be controlled by monocular light stimulation to either of the eyes. If both eyes are stimulated simultaneously the pupil reaction depeneds upon some combination of the signals originating in both eyes. Characteristic differences between the reactions to monocular and simultaneous binocular stimuli can yield information regarding the type of underlying neural interaction processes as well as the succession of neural events along the pupillomotor pathway. Qualitative comparison of the pupil reactions to monocular and binocular stimulation of sinusoidally varying intensity leads to the conclusion 1) that addition of the signals originating in both eyes occurs, and 2) that only linear transformation of the signals may take place after the addition. However, the quantitative relationship between the reactions to monocular and simultaneous binocular stimuli cannot be explained on the basis of this simple model; Data on pupil reactions to both sinusoidally modulated and flash light stimuli suggest that a mutual inhibition of the signals in the two optic nerves occurs before the addition of the signals.

Notes: Summary Comparison of the human pupillary responses to monocular and simultaneous binocular stimuli indicates that the signals evoked in both eyes by binocular stimulation first inhibit each other and then combine by addition. In this paper several possible inhibitory mechanisms are considered and a functional model is proposed which involves a shunting type non-recurrent lateral inhibition. Although the site of inhibitory interaction is not specified by the model, certain assumptions are made regarding the succession of neural events along the pupillomotor pathway. The postulated succession of transmitting stages is: nonlinear transformation, first order lowpass filter with time constant characteristic for the pupillary response, lateral inhibition, addition and second order lowpass filter with the same time constant as before. Besides predicting the experimental data this functional model resolves certain contradictions in the conclusions of different autors regarding the succession of nonlinear transformation and signal combination in the human pupillary system.

Notes: Summary The simulation of neural networks, such as the brain cortex, which have a diffuse and rather uniform structure quite unlike the simple block-structure of extant computers, leads naturally to the study of functions and principles which only in part fall within the scope of Automata Theory. Systems of decision equations must be studied with a view especially to obtaining practical means for the prevision and computation of diffuse reverberations of wanted general characteristics, with the exclusion of all others. This amounts to deriving constraints on the allowed variability of the couplings among elements during learning processes, failing which the behavior of the simulator would become uncontrollable for practical purposes. A simple mathematical treatment is presented, which essentially linearizes these problems by an appropriate use of matrix algebra and permits a straightforward study of the wanted conditions, as well as of the controlling elements which may have to be added to the network.

Notes: Summary A statistical model is presented exhibiting properties of time sequences of certain spontaneously occurring and mutually interdependent behavioural movements of a fish. The statistical description is derived from the theory of renewal processes. It is demonstrated that the probability of a movement to occur is a function only of the temporal distance to the movement that occurred immediately before. Models are developed which can be easily simulated by computer or built up in hardware technique.

Notes: Summary The lateral geniculate body performs a spatial remapping operation. This remapping may help to preserve the apparent distance of objects under symmetrical eye movements, thereby stabilizing the appearance of visual space. In addition, a related, and perhaps more basic function of the geniculate remapping may be to increase the efficiency of the neural matrix which encodes depth information. For distant fixation, the majority of cells in this cortical matrix would be responsive to crossed disparities, but for near fixation, many of these same cells might be converted into “uncrossed disparity detectors”. Two types of models for the geniculate are presented, together with supporting evidence: (1) a quantitative psychophysical model describing its steady-state properties, and (2) a qualitative neurophysiological model describing the function of the geniculate laminae.

Notes: Summary In a former paper Shannon's (undirected) transinformation content of two stochastic processes could be split in two directed parts. Here it is demonstrated how the informational exchange between single events of two processes can be measured, the so called differential transinformations adding up to the information values mentioned above. The differential transinformations are used to describe sequences of spontaneously occurring and interdependent behavioural movements of a fish. Comparison is made between correlation functions and the differential transinformations, both measuring statistical dependencies.

Notes: Summary Firstly two possible principles for the coding of messages into sequences of nerve spikes — the pulse-count-modulation and the pulse-interval-modulation — are compared; for several reasons the pulse-interval-modulation was choosen as the convenient basis for the following calculations. The information rates of afferent nerve fibers are calculated by means of data concerning the statistical properties of the spike-sequences evoced by stimulation; this data were found in the literature as well as in some own experiments. The non-linear relationship of input-signals (i.e. stimulus strength) and output-signals (mean-spike-interval = reciprocal of the firing rate) has no effect on the information rate; the influences of the channel transfer function and of the probability distribution of the input signals (the channel capacity accounts to the optimal distribution) are studied. Finally, numerical results for some kinds of afferent channels are presented; the importance of such calculations is discussed.

Notes: Zusammenfassung Eine umfassende Darstellung der Nervenzellmembran wird durch eine Kombination von Hodgkin-Huxley-Beschreibung der elektrisch erregbaren Leitwerte und Eccles-Beschreibung der synaptisch induzierten Leitwerte wiedergegeben. Diese Darstellung wird in einem elektronischen Modell veranschaulicht. Nichtlineare aktive Schaltungen werden benützt, um Leistungen zu entwickeln, die in ihrer Form mit den zeit- und spannungsabhängigen Leitwerten der Beschreibungen übereinstimmen. Die Leistungen werden mit Hilfe von Multiplikatoren in äquivalente Leitwerte umgewandelt. Das elektrische Modell enthält 24 kontrollierbare Parameter, von denen jeder mit einem in der Beschreibung übereinstimmt. Zur Einstellung der Parameter sollte man eine Strategie benützen, die soweit wie möglich die Werte von Hodgkin und Huxley (s. Tafel) als Parameter verwendet. Kleine Abweichungen von diesen Werten werden als mindere Störungen der grundliegenden Zusammenstellung betrachtet. Als Beispiel einer systematischen Untersuchung des Einflusses einer Veränderung verschiedener Parameter wurden die Spannungsschwingungen gewählt. Die Frequenz der ungedämpften Schwingungen beträgt ungefähr 50 Hertz. Verschiedene andere kleine Änderungen der Parameter können Schwingungsfrequenzen von 4 Hertz hervorrufen. Mit großen Störungen der Grundwerte kann man auch Schwingungsfrequenzen im Werte l Hertz erlangen. Es wird vorgeschlagen, daß die kombinierten Eccles- und Hodgkin-Huxley-Beschreibungen die Mehrzahl der neuroelektrischen Vorgänge einzelner Nervenzellen darstellen könnten. Dieser Vorschlag wird durch weitere Ergebnisse bestätigt. Die Beobachtung wird jedoch ausgedrückt, daß die zusammenfassende Darstellung in wenigstens einem Falle, den Nervenzellen des Hummer-Herznervenknotens, nicht zulänglich ist.

Notes: Summary The classical measure of uncertanity does not consider the observer and the special nature of the message. The aim of this paper is to introduce a measure of uncertanity, which takes into consideration both the observer and the nature of the message in such a way, that the Shannon measure and the Bongard measure are special forms of our formula.

Notes: Summary The differential equations valid for technical heat exchangers can also describe the O2 exchange in the blood capillaries and the exchange of molecules like THO and acetamid in the renal tubules. Differences in the boundary conditions occur, however. Hence, these differential equations were resolved for the corresponding boundary conditions. The results permit us to conclude that the concentration profiles occurring in the capillaries and renal tubules, as a result of diffusion in the capillary cross-section, can, generally speaking, be disregarded for the following reason: Although the differences in partial pressure between the capillary wall and capillary centre, at the beginning of the capillaries come to 40–60 mm Hg, they descrease rapidly. The calculations have shown, that the time constant for the saturation process of the plasma (10 msec), is small in comparison with the contact time of the blood (100 msec). In the tissue capillaries, the differences in partial pressure between the capillary wall and the centre come to about 4–6 mmHg. This difference remains constant over the total capillary length.

Notes: Summary The interconnection structures of the peripheral part of the nervous system, which are considered here, are two-dimensional homogeneous networks with time and space dependent inputs and outputs. The principles of connection under consideration comprise lateral inhibition and facilitation. The transfer functions of those linear networks as well as the stability problem are investigated on a digital computer using different system parameters. A closed form solution is given for an infinitely large element density which describes the network properties. In this case an inhibition system acts as high pass filter on the spatial frequencies of the input, whereas a facilitation network acts as low pass filter. The properties of the networks and the transformations in case of moving patterns are analysed using the methods of systems theory.

Notes: Summary Light quanta impinging upon the photopigments located in the rhabdomeric receptor structures of the fly's compound eyes trigger photochemical reactions which in turn elicit miniature receptor potentials (bumps). The paper mainly deals with the problem whether a single quantum of light is sufficient, or whether a coincidence of quanta and/or elementary photochemical events is necessary to trigger a miniature receptor potential. The experiments were based on tests of the optomotor responses of fixed flying flies suspended in a rotating patterned cylinder with periodic distributions of inner surface brightness. The tests were made under two different light programs: 1) Illumination constant in time 2) Illumination by periodic light pulse sequences with various frequencies. Average light fluxes absorbed by the receptors were equal in both programs. Theoretical considerations lead to the following conclusions: The strength of the optomotor responses to the light programs 1 and 2 should not differ from each other in the case of single quantum processes. However for multiquantum processes light program 2 should be more effective than light program 1 as it favours the coincidence of quantum absorptions per unit time. But these theoretical conclusions are valid only if two conditions are fulfilled in the experiments: a) The pulse frequency of light program 2 has to be kept below a certain limit which is determined by the kinetics of the photochemical systems. Otherwise light program 2 gets averaged in time and in principle can be not more effective than light program 1. b) The rates of quanta absorbed by the receptors have to be kept low enough to guarantee that the concentration of unbleached pigment molecules remains practically unchanged as compared with the concentration in darkness. Accordingly the test experiments were carried out with light pulse frequencies ranging from 500 to 1/120 cycles per second. Intensities were used which corresponded to an average quantum flux effective for one rhabdomeric structure ranging between 10 and 250 quanta per second. The interpretation of the experimental results is in accordance with the hypothesis that one single quantum of light is sufficient to trigger an elementary photochemical reaction and that in turn one single photochemical event can elicit a miniature receptor potential. At present time the experiments do not allow conclusions about the possible occurrence of coincidence-functions of synapses at the level of the first optical ganglion which receive their information via fibers leading off from the receptors. In one of the appendices of the paper, the transinformation flux into a receptor is calculated, taking into consideration the Poisson noise of the quanta disrupting the signal at extremely low quantum rates.

Notes: Summary From the communication engineers point of view the paper deals with networks consisting of linear filters and a special sort of controlled statistical pulse generators (SIG). The SIG responds to an analog signal with a sequence of Dirac impulses, where the probability for an output impulse at a given time depends on the instantaneous value of the input signal only. In connection with linear filters, however, systems can be realized in which the whole past of the input determines the statistical structure of the output. Therefore systems consisting of linear filters and SIGs (SIG networks) become interesting as models of biological systems, because in biological information processing transformations from analog signals into pulse trains occur very often. An example concerning the application of SIG systems in behavioural sciences will be discussed in a subsequent paper. In the present paper a theoretical analysis of the SIG and SIG networks is carried out by means of Statistical Communication Theory and Theory of Stochastic Processes. It is shown that under certain conditions very complex SIG networks can be treated with Correlation Theory. For one case not satisfying these conditions a solution on the base of Markoff processes is given.

Notes: Summary As an extention of the law of activity (Sato, 1968), the law of interactivity was established by which new physiological meanings are added to such well known basic concepts as facilitation and occlusion in the central nervous system, and super- and subnormal phases in the recovery function of physiological systems.

Notes: Summary Transmission of nerve signals in the crayfish brain was studied by means of the transfer function derived from input-output analysis with random stimulation. The transfer function was measured in the form of frequency-response-function and represented by Bode plot. Two classes of the frequency-response-functions were discriminated, corresponding to two types of response patterns evoked by the constant frequency stimuli. The first type had the characteristics of a band pass filter similar to an underdamped resonant circuit. The second one had the characteristics of a phase lag circuit, occasionally, with additional small positive or negative peak at almost the same frequency as the resonance of the first type. A few possibilities for the resonance and the phase lag mechanisms were discussed.

Notes: Abstract Slow electrical responses were recorded from receptors and from the lamina of the visual pathway of the fly Musca. (a) Receptors 1 to 6 in the retinal ommatidia are identified by their response dichroic sensitivity planes. The half-width of their angular sensitivity distributions is estimated 2.5° in dark adaptation, and found not to vary with ambient illumination. The retinula cells are only excited by light that enters the eye through their overlying corneal facets. (b) The responses of the lamina show no detectable dichroic sensitivity, though in favourable cases their angular sensitivity distributions may be as narrow as those of the receptors. It is shown that these responses are excited by light that enters the six facets of the corneal projection of the single lamina cartridge synapse. The retinula fibres of passage through the lamina, originating from ommatidial cells 7 and 8, evidently do not contribute excitation to the responses. (c) It is shown that the separate responses contributed by the individual receptors of the projection are added linearly at the lamina response compartment over a wide range of light intensities.

Notes: Abstract In the field of control, it is often necessary to classify a large set of different input vectors into a small number of different classes. The input vectors which for example represent the state of a controlled system, define points in the so-called “pattern space”, and the classification may be regarded as a separation of these points. The structures of two different trainable classifiers for linear and piecewise linear separation are described: the madaline and the learning-matrix. As a training method for the learning-matrix with minimum distance classification, two algorithms are introduced: the method of weighted shifting and the method of error projection.

Notes: Abstract In investigating the response of systems to random input events, “dead times” in registering these events are met with, as in the case of neuronal behaviour. These situations are studied in terms of product densities making use of the renewal nature of the problem. Different types of cumulative responses of systems are investigated. Some interesting features of a system, which breaks down at a critical value of the cumulative response are analysed.

Notes: Summary Computer simulation of a relatively simple model can reproduce the main characteristics of the firing patters of some nerve cells. The abdominal stretch receptor of the crayfish has provided an analogous biological model. Synaptic input impulses were simulated by randomly distributed, short lasting transmembrane current pulses. Under these experimental conditions the stretch receptor neurone largely behaved as predicted by the computer simulations.

Notes: Zusammenfassung Die Frage wird aufgeworfen, inwiefern man aus der Statistik der Impulsfolgen bei Spontanaktivität einzelner Neurone auf die Funktionscharakteristik des Neurons selbst und auf die Art seiner Einschaltung in das Nervennetz schließen kann. Die Untersuchung der Verteilung der Intervalle verschiedener Dauer zwischen aufeinanderfolgenden Aktionspotentialen von Purkinjezellen des Froschkleinhirns ergibt, in Übereinstimmung mit den von anderen Autoren bei anderen Neuronentypen erhobenen Befunden, zwei Arten von Histogrammen: einerseits unimodale, vorwiegend bei Neuronen mit höherer Impulsfrequenz, andererseits bi- und trimodale, vorwiegend bei niedriger Impulsfrequenz. Die Form der unimodalen Verteilungen läßt, unter Annahme einer Gaußschen Verteilung der Erregungsniveaus im Eingang des Neurons, die sich aus der Summation einer großen Anzahl voneinander unabhängiger afferenter Erregungen ergibt, auf einen nichtlinearen Zusammenhang zwischen Erregungsniveaus und Dauer der Intervalle schließen; die unsymmetrische Verteilung der Intervalle wäre demnach als eine Verzerrung der statistisch gegebenen symmetrischen Verteilung der Erregungsniveaus zu verstehen. Diese Erklärung hat den Vorteil, kontinuierlich zu den bei niedriger Impulsfrequenz beobachteten unregelmäßigen Verteilungen überzuleiten, die sich einfach aus dem Zusammenbruch der statistischen Regelmäßigkeit im Falle von einer geringen Anzahl aktiver Fasern im Eingang ergeben. Die Untersuchung von Korrelationen innerhalb der Impulsfolgen mit Hilfe einer Verzögerungs- und Koinzidenzschaltung ergibt scharfe Maxima der Impulswahrscheinlichkeit als Funktion des zeitlichen Abstands von vorhergehenden Impulsen. Da diese Autokorrelogramme bei verschiedenen Neuronen derselben Art (Purkinjezellen des Froschkleinhirns) verschiedene Verläufe zeigen, wird zu ihrer Erklärung weniger eine Eigenschaft des Einzelneurons (Schwellenänderung nach em Aktionspotential) herangezogen, als die Rückwirkung des eurons auf sich selbst auf dem Umweg über andere Neurone es Nervennetzes. Die Korrelation zwischen Impulsreihen in erschiedenen Neuronen desselben Nervennetzes, die auf Grund ieser Annahme zu erwarten ist, wird in einigen Beispielen uch gefunden, wovon eines im Anhang gezeigt wird.

Notes: Zusammenfassung Am Subcoxalgelenk befinden sich außer den schon bekannten Borstenfeldern Proprioreceptoren in Form von vier Borstenreihen an der Coxa. -Die Bewegung des Femur-Tibia-Gelenkes wird von einem Chordotonalorgan gemessen, das an der Basis des Femur liegt. Vom Receptor zieht eine cuticulare Sehne (Receptorsehne) zum FemurTibia-Gelenk. Die wichtigsten Nervenverästelungen im Femur und eine anormale Lage des Chordotonalorganes werden beschrieben. -Das Chordotonalorgan ist Glied eines Regelkreises zur Stabilisierung des Femur-Tibia-Gelenkes. Dieser Regelkreis adaptiert, mindestens bei höherer Belastung, langsam, aber vollständig. —Wirkt bei einem senkrecht vom Körper abstehenden Bein eine Kraft in Richtung der Querachse auf das Tier ein, ist in der normalen Körperhaltung die Auslenkung des Tibia-Tarsus-Gelenkes für kurze Zeit proportional zur einwirkenden Kraft. Die Regelkreise der beiden Körperseiten beeinflussen sich nicht gegenseitig. —Die von der Streckmuskulatur erzeugte Kraft ist um so größer, je stärker der Receptor vor Beginn des Reizes gedehnt war. — Wird die Receptorsehne nach außen gezogen, streckt das Tier das Femur-Tibia-Gelenk. Wird die Receptorsehne nach innen geschoben, beugt es das Femur-Tibia-Gelenk. Dabei ist ebenfalls vollständige Adaptation zu beobachten. — Die Streckung der Tibia (in Winkelgraden) ist proportional dem Logarithmus der Bewegung der Receptorsehne nach außen. Die Reaktion ist um so stärker, je mehr der Receptor vor Beginn des Reizes gedehnt war. —Die Beugung der Tibia (in Winkelgraden) ist proportional dem Logarithmus der Bewegung der Receptorsehne nach innen. Auch diese Reaktion ist um so stärker, je mehr der Receptor vor Beginn des Reizes gedehnt war. —Wird eine senkrechte Lauffläche von der Seite beleuchtet, stellen sich die Tiere teils in eine Resultierende zwischen Licht-und Schwerkraftrichtung ein, teils wenden sie sich vom Licht ab. — Der Mittelwert der Winkel zwischen Tierlängsachse und Schwerelot (α1) ist bei den dem Licht zugekehrten Tierstellungen von der Lichtintensität und dem Winkel zwischen Lichtrichtung und Schwerelot abhängig. Er ist unabhängig von Körpergewicht und Hangneigung. Die Streuung wird bei erhöhtem Körpergewicht kleiner. Abschaben der Sinnesborsten an den Subcoxalgelenken verkleinert den Mittelwert der Winkel α1. Werden die Sehnen der femoralen Chordotonalorgane der nach oben zeigenden Körperseite durchtrennt, wird der Mittelwert der Winkel α1 kleiner. Bei derartig operierten Tieren wird der Mittelwert der Winkel α1 nach Erhöhung des Körpergewichtes größer. Werden die Sehnen der femoralen Chordotonalorgane der nach unten zeigenden Körperseite durchtrennt, wird der Mittelwert der Winkel α1 größer als bei intakten Tieren. Bei derartig operierten Tieren wird der Mittelwert der Winkel α1 nach Erhöhung des Körpergewichtes wieder kleiner. — Werden die Sehnen der femoralen Chordotonalorgane einer Körperseite durchtrennt, weichen die Tiere auf einer senkrechten Fläche zur operierten Körperseite hin von der Senkrechten ab (intakte Tiere laufen unter denselben Bedingungen etwa senkrecht nach oben oder unten). Der Winkel zwischen Körperlängsachse und Schwerelot ist bei den operierten Tieren um so kleiner, je größer das Körpergewicht und je größer die Hangneigung ist. — Die Genauigkeit, mit der ein einmal eingeschlagener Kurs nach Drehung der Lauffläche wieder aufgenommen wird, ist um so größer, je steiler die Lauffläche steht. — Bei der Orientierung im Schwerefeld liegt die Labilit ätsstellung für die Stabilitätsstellungen 0° und 180° ungefähr gegenüber der jeweiligen Stabilitätsstellung. — Es wird festgestellt, das Tier verhalte sich in allen Experimenten so, wie wenn bei ihm die von der negativen Geotaxis ausgelöste Drehtendenz als Quotient aus der Belastung in Richtung der Querachse und dem Betrag der Belastung in Richtung der Längsachse gebildet würde. Ein Minimalmodell für die Bildung der Drehtendenz wird aufgestellt. Theoretisch denkbare Möglichkeiten zur Verschiebung der Stabilitäts-und Labilitätsstellung werden diskutiert.

Notes: Summary The aim is to design a machine which is able to learn a number of idealised characters and to recognise them, irrespective of their size, position and context on an infinite retina. If the number of characters which such a machine can possibly learn to recognise is astronomical, it is not practical to use separate templates for every possible character. It is more economical to use, instead, templates for various parts, called features, of characters. In recognising a number of characters simultaneously, without scanning, the question arises of how to tell which feature belongs to which character of figure on the retina. In particular, if a given character is not present but all its features are included in nonsense figures simultaneously present on the retina then the machine must not indicate the presence of the given character. The technique which overcomes this difficulty employs overlapping features which must be mutually consistent for recognition. This consistency technique is assessed by comparison with a more conventional technique, and the work is restricted to closed line characters which are not subject to deformations or mutilations.

Notes: Summary At first an intuitive interpretation of the two fundamental terms of information theory, namely information and transinformation, is given. Then the physical limits for information transmission are shown and the problem of how much energy is necessary in order to transmit one bit is examined. In the following part biological limits concerning detection and processing of information by man are discussed.

Notes: Summary A simulation study is reported of the spread of excitation in a digital computer model based quite realistically on a coelenterate nerve net. The question posed is whether an elementary nervous system with randomly distributed properties can discriminate between time patterns of stimuli at the same average frequency. Forty-four temporal patterns of stimulation, each composed of seven stimuli in the same total period of time were applied to each of nine simulated nerve nets with each of eleven different distributions of four rates of decay of facilitation. The results may be summarized as follows: 1. The simulated nerve nets used were able reliably to discriminate between many of the time patterns used. The factors entering into range and acuity of pattern discrimination by the net are identified. 2. The simulated nerve nets tended to support a greater spread of excitation in response to even temporal distribution of stimuli than in response to clustered distributions of stimuli under certain conditions. These conditions are specified. 3. The response measure which was used (“final spread”) is shown to give different results in some cases than either of two other measures (“average” spread for the seven stimuli and “maximum spread”). 4. The simulated nerve nets were able to produce reliable differences in the spread of excitation between certain patterns and their temporally mirror-image counterparts. The necessary conditions for such pattern recognition are described. The relationships among the principal variables, namely the temporal distribution of stimuli, the specific sequence of junction decay rates, the magnitudes of facilitation decay rates, and their relative frequency distributions are described as they affect the spread of excitation in the nerve net. The overall finding that the net is able to discriminate between some temporal patterns gathers significance in that it represents an ability of the net to translate temporally coded information into spatial form. Thus, it is shown that already at the level of a simplified model of a coelenterate nervous system, the requisites for temporal to spatial translation are met.

Notes: Summary Studies of the optomotor response, the tendency to turn in response to a moving pattern, have yielded some understanding of the motion detection capabilities of the fly. We present data from extracellular microelectrode recordings from the optic lobes of the housefly, Musca domestica and the blowflies Eucalliphora lilaea and Calliphora phaenicia. Directionally selective and directionally nonselective motion sensitive units were observed in the region between the medulla and the lobula of all three species. Employing similar stimulus conditions to those used in the optomotor reaction studies, it was found that the response of the directionally selective units exhibited most of the characteristics of the optomotor response torque measurements. It is concluded that these units code the information prerequisite to the optomotor response and hence, that much data processing is achieved in the first few synaptic layers of the insect visual nervous system.

Notes: Zusammenfassung Die Netzhaut decerebrierter Katzen wurde mit sinusförmig moduliertem Licht gereizt und die in den Ganglienzellen ausgelöste Erregung extracellulär registriert. Amplitude und momentane Frequenz der Aktionspotentiale ändern sich sinusförmig und besitzen zueinander eine Phasenverschiebung von 180°. Die Phasenverschiebung ist unabhängig von der Frequenz des Reizlichtes, die im Bereich von 0,1–10 Hz geändert wurde. Anhand von Kontrollmessungen wurde gezeigt, daß die Amplitudenänderung der gemessenen Aktionspotentiale auf Änderungen des Membranpotentials beruht.

Notes: Summary Evidence is presented for a simple model of the visual system, assuming that information processing is performed by Fechner type responses at all levels of neural interconnections. Red-green, yellow-blue and white-black opponentcolor responses are built up at the retinal level by linear algebraic summations of the logarithms of the excitation energies, that are produced in red, green and blue sensitive photoreceptors. Spacetime correlations are obtained by further nearly linear algebraic summations of the opponent-field responses in separate opponent-color response channels. This model provides a good first approximation for the description of various color vision phenomena, as the perception of brilliance and chromaticity of single colors, brilliance contrast, color induction, and Land's two color projections.

Notes: Zusammenfassung Mechanisierte Sprachübersetzung, Speichern von Information und Mustererkennung stellen drei Sonderfälle der allgemeinen Operationen „Lernen”, „Ordnen” und „Erkennen” dar. Die Automatisierung solcher Operationen bildet nicht nur ein an und für sich interessantes Problem, sondern eine dringliche Notwendigkeit, sollen die universellen Gro Brechenanlagen nicht an der Komplexität ihrer inneren Organisation (soft-ware) ersticken. Als vielleicht einfachstes dieser drei — im Grunde nie getrennt auftretenden — Vorgänge, wird das Problem der Wiedererkennung betrachtet, d.h. dasjenige der Abbildung eines kontinuierlichen Ensembles reeller, metrischer Objekte auf eine endliche, diskrete Struktur ohne Metrik, wie sie beispielsweise das „Innere” einer digitalen Rechenanlage darstellt. Der notwendige Übergang vom heute einzig üblichen Speicher zum eigentlichen Gedächtnis hängt unter anderem von der Lösung dieses Problems ab. Seine Untersuchung hebt die grundsätzliche Bedeutung der Umgebung einzelner Objekte — ihres Kontextes — hervor, welche Umgebung im diskreten Raum das notwendige Äquivalent zum Kontinuum darstellt. Dies führt zu einer neuen Betrachtung des Paradoxon von Brillouin, und zur notwendigen Benützung dynamischer Methoden bei der Verarbeitung von Information, anders gesagt bei ihrer Überführung aus der bildhaft-kontinuierlichen Form in die diskrete, und zurück. Womit zumindest eine mögliche Definition nicht-trivialer Informationsverarbeitung gegeben ist.

Notes: Summary The exact mathematical treatment is given for a non linear equation describing the delayed yes-or-no response to a binary system to external stimulations, in some typical cases of interest. Comparison is made with neurophysiological data on the frequency rate of firings of stimulated neurons; the same equation, however, can be conceivably applied to a vast variety of phenomena. The procedures followed to solve the problems that arise in connection with this equation could be extended to more general types of non linear equations.

Notes: Summary This paper is concerned with a model for the interaction of certain behaviour patterns of a fish (Pelmatochromis subocellatus kribensis BOUL; Cichlidae). Five behaviour patterns (schooling behaviour, locomotion, biting, fleeing, camouflaging) are described as a function of four external stimulus patterns (CO2-concentration, school-signal, conspecific's signal, frightening-signal) and three so-called internal states of readiness (readiness to school, readiness to attack with bites, responsiveness to frightening stimuli). In accordance with the statistical appearance of behaviour patterns the model contains five interdependent statistical pulse generators. This model is planned to find and to confirm a theory for a quantitative description of behaviour patterns, the interdependencies of which give a representation of the internal structure of the organism.

Notes: Summary The static regulation factor of the pupil servosystem in man, as dependent on the intensity of the light-stimulus, was calculated. Data were collected from own experiments, in which the whole field of view of one eye was homogeneously illuminated by white light. For these calculations the author developped a suitable definition of the regulation factor, using ordinary and partial differential quotients. The optimum of the regulation factor was found to be 2/3 at a luminance of 200 abs, i.e., at this point 1/3 of the slight variations in light intensity are compensated.

Notes: Summary This paper presents a new conception of information theory, which proved to be a generalisation of Shannon's information channel in the sense of a bidirectional communication. Consequently both, transmitter and receiver, are to be stochastic information generators. Transinformation is the result of a stochastic synchronisation process of one generator by the other, which process may occur in both directions. However the sum of both degrees of synchronisation is found to be limited to equal or less than one. Mathematically the theory describes the mutual statistical influence of two stochastic processes. Application to a quantitative description of communication between human beings is envisaged. Due to the fact that conscious processes are individually different, the corresponding information quantities are called “subjective.”

Notes: Summary In the central nervous system information transmission and processing are accomplished by pulses and pulse trains. The superposition of pulse trains is essential for information processing as it allows the execution of several logical operations, e.g. the multiplication of afferent signals. Jenik (1961) has pointed out that for the superposition of periodic pulse trains the rate of coincidence is proportional to the product of the pulse repetition frequencies (“multiplication law“). Furtheron he has shown that this simple principle is not always applicable. Errors may occur for certain repetition frequencies of the pulse trains. If the product of signals is accomplished by superposition, mechanisms must exist reducing these errors. Since pulse trains in the nervous system always vary stochastically the following paper is concerned with the effect of random pulse trains in superposition. Two different types of random pulse trains are investigated, trains with random phase and trains with random interval between the pulses. For these two cases calculation methods are given. It is also shown that the deviations from the “multiplication law” may disappear when superimposing random pulse trains.

Notes: Summary The formation of the interconnection stimulus-response in a learning system is analysed. The system, a technical or a biological one establishes this correspondence by processing the information fed back from the medium during the learning process. This information has two aspects: a quantitative one related to the probability of the events, and a qualitative one related to the utility of the events in view of a goal. Both aspects are taken into consideration; by successive experiences the system eliminates the double uncertainty concerning the probabilistic dependence: stimulu — sresponse — outcome and its utility. Learning implies then a system for evaluating the utilities of different outcomes in view of a goal, a memory to record them and a decision system for selecting the corresponding responses upon a given criterion.

Notes: Zusammenfassung Sinusförmig wechselnde Kräfte werden am Auge angewandt, und die resultierenden Bewegungen werden mit Hilfe eines Beschleunigungsmessers gemessen. Dieser ist an einer Kontaktlinse befestigt. Die Veränderung der Größe und des Phasenwinkels der Augendrehung mit der Frequenz wird in Form eines Bode-Diagramms wiedergegeben. Ein mechanisches Modell, das aus linearen visco-elastischen Elementen besteht, wird verwendet, um das Augapfel-Muskel-System nachzuahmen. Die Parameter des Modells werden nach der Methode der besten Übereinstimmung aus den Übergangs-Charakteristika bestimmt. Die Beschleunigung-gegen-Zeit Kurve der Bewegung, die durch Anwendung einer stufenartigen Verdrehung am Auge verursacht wird, ist in guter Übereinstimmung mit der vom Modell vorausgesagten. Die Drehwinkel-Übergangsfunktion des unbelasteten Auges wird aus dem Modell durch Abzug des Trägheitsmoments der Kontaktlinse und ihrer Anhängsel abgeleitet. Ergebnisse für Horizontal- und Vertikalbewegungen werden gesondert diskutiert. Vier kanonische und eine nicht-kanonische Form des mechanischen Modells sind angegeben. Angesichts der bekannten mechanischen Eigenschaften freiwilliger Muskel wird vermutet, daß die nicht-kanonische Form am ehesten physikalischen Elementen in der Augenhöhle entspricht. Drei Arten der Augenbewegung, die für das Sehen wichtig sind, werden verglichen mit der Mechanik des Augapfel-Muskel-Systems. Es wird vorausgesagt, daß eine Rotationsresonanz des Augapfels in der Pfanne erzeugt werden kann, wenn der Kopf in geeigneter Weise in Schwingungen gebracht wird. Die Natur der Muskelkräfte, die für saccadische Bewegungen und unwillkürliche Fixierungstremore verantwortlich sind, wird aufgeklärt.

Notes: Abstract Fig. 4 demonstrates the problem: How are the messages of the cones processed by the structures of the CNS before they are perceived in the form of colour impressions? The absorption curves of Fig. 2 are used as input functions of the three types of cones; they were adapted from the work of MacNichol et al. [2] and from similar results of other authors [1, 3, 4]. According to a well known hypothesis [8] the messages of the green receptors are supposed to be the antagonists to the messages of the “red” receptors, and the messages of the blue receptors the antagonists to the sum of the messages of the green and the “red” receptors (Fig. 11). But this hypothesis does not account for the red sensations present in the region of both the long and the short wave end of the visual spectrum (Fig. 14). Therefore another model was studied (Fig. 12): Blue (P 445, cyanolabe) being antagonistic to yellow (P 570, erythrolabe); and green (P 535, chlorolabe) being antagonistic to the combined action of blue and yellow (P 445 and P 570). This formalism explains the following experimental data: The distribution of colour sensations along the spectrum (Fig. 7, 13, 14); the wavelengths of antagonistic spectral colours (Fig. 5); the general form of the spectral hue discrimination function (Fig. 8); the number and the qualitative symptoms of colour deficiencies (Fig. 10, 18–24). A meaningful interpretation is given concerning the reduced colour vision of the distal parts of the retina.

Notes: Zusammenfassung Mit Hilfe eines Digitalrechners wurden die Eingangs-und Ausgangsbeziehungen auf synaptischer Ebene untersucht und dargestellt. Diese Untersuchung erstreckt sich auf erregende Synapsen und analysiert die Veränderungen postsynaptischer Aktionspotentiale, die auftreten, 1. wenn die Anzahl der präsynaptischen axonischen Endigungen ansteigt, während andererseits die Amplitude des EPSP abnimmt; 2. wenn sich die statistische Struktur oder „Form” der Spike-Kette in jeder präsynaptischen Faser verändert; und/oder 3. wenn die Beziehungen zwischen den präsynaptischen Fasern von völliger Unabhängigkeit bis zu starker Abhängigkeit variiert werden. 1Unabhängige präsynaptische Endigungen. Mit zunehmender Anzahl präsynaptischer Endigungen bei gleichzeitiger proportionaler Abnahme des EPSP (Input Form konstant) treten folgende Veränderungen auf: a) das durchschnittliche Intervall zwischen den postsynaptischen Spikes nimmt etwas zu; b) die mittlere statistische Abweichung (standard deviation) nimmt erheblich ab; c) die Form des Histogramms wird eng umschrieben; und d) das Autokorrelogramm nimmt „periodischen” Charakter an. Wenn andererseits die präsynaptische Form verändert wird (während die Anzahl der Endigungen sowie die Größe des EPSP konstant bleibt), hängt der am postsynaptischen Ausgang registrierte Effekt von der gegebenen Anzahl der Endigungen und von der Größe des EPSP ab. Ist die Anzahl der Endigungen gering und das EPSP groß, dann variiert der Ausgang mit der präsynaptischen Form. Der postsynaptische Variationskoeffizient steht dann in linearer Abhängigkeit vom präsynaptischen Variationskoeffizienten, wobei die Steigung der Geraden mit zunehmendem Eingang abnimmt. Sind die Endigungen zahlreich und die Größen der EPSPs gering, dann übt die präsynaptische Form keinen Einfluß mehr aus, und das von der postsynaptischen Zelle erzeugte Ausgangsprodukt wird unabhängig von der detaillierten Struktur des Eingangs. Für eine jegliche Kombination von unabhängigen und schwachen Eingangsformen stellt sich das Ausgangsprodukt in Form einer sehr regelmäßig gestalteten und durch gleichmäßige Abstände gekennzeichneten Spike-Kette dar (diese Folgerung gilt nur für die Fälle, in denen die präsynaptischen Endigungen sich nicht äußerst langsam entladen). Diese Resultate können mathematisch an Hand eines einfachen Membranmodells abgeleitet werden (s. Appendix). Wenn das EPSP in Größe ansteigt, alle anderen Variablen jedoch gleich bleiben, dann nimmt das postsynaptische Intervall fortwährend ab. Dieser Abfall ist entweder gleichmäßig (präsynaptische Form: „Poisson”) oder stufenweise (präsynaptische Form: „pacemaker”). Die postsynaptischen Aktionspotentiale werden durch eine vergleichsweise kleine Anzahl von hemmenden Endigungen wirkungsvoll blockiert. 2Abhängige präsynaptische Endigungen. Wenn sich der Grad der Abhängigkeit zwischen den präsynaptischen Endigungen in physiologischen Grenzen bewegt, dann kann die Aktivität der postsynaptischen Zelle als eine Funktion der statistischen Form der Eingangskanäle angegeben werden, und das sogar, wenn die letzteren zahlreich und schwach sind. Dieser Fall tritt dann ein, wenn die Abhängigkeit zwischen den präsynaptischen Endigungen nur einen Teil aller Endigungen oder nur die Endigungen innerhalb getrennter und unabhängiger Gruppen betrifft. Um die im Nervensystem stattfindenden Übertragungen zu verstehen, ist es notwendig, diejenigen präsynaptischen Statistiken zu idendifizieren, die entsprechende postsynaptische Entladungen beeinflussen. Wenn präsynaptische Endigungen große PSPs hervorrufen, dann ist ihr Einfluß dominierend und wird entsprechend der präzisen statistischen Form ausgeübt. Wenn die Endigungen kleine PSPs hervorrufen, dann hängt ihr Einfluß weitgehend von dem Grad der Abhängigkeit voneinander ab. Wenn die präsynaptischen Endigungen unkorreliert sind, dann vermitteln ihre durchschnittlichen, präsynaptischen Entladungsgeschwindigkeiten eine gleichmäßige Regulierung des postsynaptischen Membranpotentials und der postsynaptischen Entladungsgeschwindigkeiten. Sind andererseits die präsynaptischen Endigungen korreliert, dann nehmen sie eine dominierende Funktion ein, und die Beziehungen zwischen präsynaptischer Regulierung und postsynaptischer Entladung können präzise definiert werden. Somit stellt sich der Grad der Abhängigkeit zwischen den präsynaptischen Endigungen als eine funktionell bedeutende Variable dar.

Notes: Summary Optical characteristics of the dioptric system in the ommatidia of Musca have been analysed by use of “antidromic illumination” of the eye. The results indicate that the distal endings of the rhabdomers terminate near the focal plane of the dioptrics and that the quality of the lens is high enough to resolve some details of their shape. — Using optical methods it has been possible to confirm directly that the optical axes of 7 individual rhabdomers from 7 different ommatidia all converge to a common point in the distant surroundings. This is a characteristic for compound eyes of the “neural superposition” type. — The results are discussed on the basis of the hypothesis that the Musca eye is composed of two functionally different subsystems: One system (D) with high absolute sensitivity and low spatial resolution consisting of the sense cells no. 1 to 6, and a second system (H) with high spatial resolution and low absolute sensitivity composed of cells no. 7 and 8.

Notes: Summary A function model is proposed describing the perception strength of vibration as a function of the stimulation strength and its just noticable change by the sense of touch. The stimulus consists of a sinusoidal deformation of the skin with any variation of the vibration amplitude in time. The model is based on results reported in other papers [1–3]. Considering only one stimulation place the function model consists of a chain connexion of links with simple transmission functions like square, proportional, differential, integration function and others. If several stimulation places are allowed simultaneously, the function model must be completed by a variable “Ortsfilter”. In spite of the simplicity of the function model the calculated values show a good conformity with the measured values.

Notes: Summary A uniform block diagram has been designed describing the normal cell increase as well as that of goodnatured and malignant tumours. It was shown that the various cancer diseases have their origin in changements of the structure data of the controller. Frequency and transition responses, locus diagrams, and stability conditions of the above mentioned cases were studied. The stability determination has shown that the instability of the automatic control loop for the cell increase in malignant tumours can be made stable by a chemotherapeuticum to be developed influencing the molecular structure of the regulator genes so that the frequency response of the controller regains its positive component in the portional band, which was lost in the course of the disease.

Notes: Summary By modulating sinusoidally the impulse frequency of the inhibitory input to the slowly adapting stretch receptor of Crustacea, the impulse frequency of the sensory neuron becomes sinusoidally modulated. The systems involved are linear for amplitudes of modulation up to 70% of the steady state frequency. The transfer function is given and some of its implications are discussed.

Notes: Abstract The aim of the present paper is to present some applications of the information theory to the problem of the algorithms of recognition. In all these considerations, a qualitative quantitative measure of the information is used to give a generalisation of Landa's algorithm of recognition.

Notes: Abstract A model of memory incorporating several unusual features is discussed. It is assumed that stored traces are added together in a common storage array. When the recovery system is presented with an input pattern, three reasonable demands may be made of the storage array for the memory system to be useful: has anything similar to the input pattern been stored previously (recognition); what are the significant differences between the input pattern and the stored pattern (reconstruction); what other stored patterns are similar to the input pattern (association). Methods of meeting these demands are discussed. A numerical example embodying these concepts is presented.

Notes: Summary Psychoacoustic experiments are reported in which the subjects had to state which one of two aural sensations had been varied predominantly. The sensations investigated are loudness, pitch and roughness of amplitude modulated tones. Proper analysis of the subjects' responses results in those intervals of these three sensations that are equivalent to each other. The results can be uniformly described by a simple law which means that equivalent intervals of sensations include the same number of just noticeable differences of each sensation.

Notes: Summary Evoked potentials were recorded from the hypothalamus of acutely-prepared estrous guinea-pigs responding to electrical stimulation of the vaginal cervix. The latency ranged from 38 to 60 msec. The peak of the first positive wave appeared 50 to 90 msec after the stimulation, usually followed by a negative wave in 120 msec. These responses were obtained from the mid-hypothalamus. The cervical stimulation subsequently induced the release of hypophysial luteotropin (LTH) as indicated by the occurrence of pseudopregnancy. According to the results, the mechanism triggering hypothalamic potential changes was discussed in connection with the LTH secretion.

Notes: Summary Discharges of individual fibers from the IX-th frog dorsal root, under stretching of the hind leg by falling weights, have been recorded. About 1/3 of the recorded fibers are leading off from rapidly adapting receptors. The fibers coming from slowly adapting receptors and those not responding to the applied stimulus are analyzed with some details. The number of pulse per second as a function of time and load is examined. It has been found that a group of fibers increases the rate of firing regularly with the load, a second group gives the greatest response in intermediate zones of the loads scale and a third group shows a higher threshold. Some remarks on the aggregate message reaching the spinal cord are developed.

Notes: Zusammenfassung Als bestimmend für die Dynamik des Langzeitvorganges des Spannungsabfalles im Muskel nach aufgebrachten stufen- und rampenförmigen Längenänderungen ist das Übertragungsverhalten eines s k -Elementes mit einer elastischen Komponente ermittelt worden. Vergleichende Betrachtungen über die Veränderung der H-Zonenlänge und des Exponenten k haben zu einer Lokalisierung der Übertragungseigenschaften in der in der Mitte des Sarkomers befindlichen H-Zone geführt. Es wird vorgeschlagen, daß die elastische Rückstellkraft des Muskels in den parallel zur H-Zone befindlichen m-Filamenten generiert wird; diese Strukturen stellen zusätzliche Verbindungen zwischen den Myosinfilamenten in den beiden Halbsarkomeren dar. Aus der Modellierung mit einem Analogrechner folgt, daß das s k -Element physikalisch einer Kombination einer stark nichtlinearen Feder mit einem konventionellen Dämpfertopf entspricht. Durch vektorielle Subtraktion des s k -Anteiles mit elastischer Komponente von der Ortskurve der Dehnbarkeit des entspannten Insektenflugmuskels ist mit sehr guter Genauigkeit die Übertragungsfunktion von zwei in Serie befindlichen viskoelastischen Elementen des Maxwell-Typs ermittelt worden. Durch die Annahme, daß in einem der drei bereits früher postulierten Maxwell-Elemente, die als konzentriert in den dominanten passiven Strukturen — den Verbindungsfilamenten, den Myosinfilamenten und der H-Zone — angenommen wurden, die Feder in Serie mit dem Dämpfertopf eine stark nichtlineare Kraft-Längen-Charakteristik hat, wird es möglich, die Gültigkeit eines schon früher formulierten Modelles auf Langzeitveränderungen ausdehnen. Die Möglichkeit der Beschreibung von Spannungsabfällen, gemessen für verschiedene andere Muskeln, durch die Übertragungseigenschaften eines s k -Gliedes läßt eine Allgemeingültigkeit der am Insektenmuskel erzielten Ergebnisse als wahrscheinlich erscheinen. Die Möglichkeit ist diskutiert, daß die für Dehnungsreceptoren abgeleitete Übertragungsfunktion das im Muskel lokalisierte s k -Glied mit einbezieht.

Notes: Zusammenfassung Das Gefäßsystem isolierter Nieren und isolierter Herzen von Ratten wurde mit einer zellfreien, sauerstoffgesättigten isotonischen Nährlösung künstlich durchströmt. Dabei wurden a) die Volumenstromantworten (als Ausgangssignal) auf rechteckförmige Drucksprünge (als Eingangssignal) analysiert und b) die Volumenstrom-Druckbeziehung (V/P) isolierter Nieren bei sinusförmig moduliertem Eingangs(Druck-) signal und bei unmoduliertem (DC) Eingangssignal mit dem Ziel untersucht, eine aufgrund früherer Untersuchungen vermutete (Druck-)steilheitsabhängige Reaktionskomponente nachzuweisen (Başar et al., 1968a). 1. Bei gleichem mittleren Perfusionsdruck liegt der mittlere Perfusionsvolumenstrom im Druckbereich der sog. Autoregulation (zwischen 100 und 220 mm Hg) bei sinusförmig moduliertem Druck signifikant niedriger als bei Perfusion mit Gleichdruck. 2. Volumenstromantwortkurven auf Aufwärtsdrucksprünge (von 90 auf 180 mm Hg) enthalten eine differentialquotientenempfindliche Antwortkomponente, während die Registrierung von Abwärtssprungantworten (180 auf 90 mm Hg) ein proportionales Verhalten ergibt. Die Ergebnisse werden im Licht des kürzlich von Clynes (1961, 1962) erarbeiteten Konzeptes einer „unidirectional rate sensitivity” diskutiert. Übertragungsfunktionen, die den Zusammenhang zwischen dem Perfusionsdruck und der resultierenden Perfusionsstromstärke in der autoregulierenden Niere und dem Coronargefäßsystem des Herzens beschreiben, werden aus Bode-Diagrammen abgeleitet und miteinander verglichen.

Notes: Abstract The evidence of averaged evoked potentials has been proved. The resulting possibilities for calculation and reduction of the standard deviation are discussed with special regard to computer application. Examples of acoustic evoked responses from the cortex of the cat and the scalp of man demonstrate the suggested processing of EEG data.

Notes: Summary It is shown that sets of quantities that are investigated in hearing-experiments, namely sounds and sound-sensations, can be referred to as input- and output-sets of conciously percepting systems. A simple system-theoretic model of a conciously percepting system is given, the set-functions between input and outputs are derived. The basic steps to measure these set-functions and to estimate accompaning possible errors are discussed. The results can be generalized on perceptional experiments concerning other senses than hearing.

Notes: Summary Physical measurements of the energy content of effective optomotor stimuli for the fly Musca support the notion that single light quanta trapped by the photopigment of the retinal receptors generate adequate sensory messages. The results of an electrophysiological experiment on single retinula cells in the fly's eye are presented to strengthen confidence in certain assumptions it was necessary to adopt with regard to receptor optical and spectral properties in order to calibrate the complex movement stimuli. This experiment also emphasizes the small amplitude of electrical signals generated in the receptors by dim but behaviourally effective pattern illumination levels.

Notes: Zusammenfassung Die Formeln zur Berechnung der Periodentonhöhe („Tonhöhe des Residuums”) aus den für sie wichtigsten Schallgrößen, der Hüllkurvenperiode des Signals und der Frequenz des untersten Teiltons des Gemisches, liefern ein Funktionsschema, gemäß welchem die Periodentonhöhe die physikalische Subharmonische dieses Tons ist und außerdem der Hüllkurvenfrequenz möglichst nahe kommt. Ein elektrisches Modell, das diese Funktionen simuliert, liefert Ergebnisse für die Abhängigkeit der Periodentonhöhe von den beiden Schallgrößen, die mit den Meßergebnissen gut übereinstimmen. Um auch die Abhängigkeit der Periodentonhöhe von anderen Änderungen des Schallreizes zu beschreiben, z.B. vom Pegel des Schalles oder von dessen pauschaler Frequenzlage, wurde dieses Funktionsschema so verbessert, daß die Periodentonhöhe als empfundene Subharmonische der Tonhöhenempfindung des untersten Teiltons anzusehen ist. Mit einem solchen Schema kann man mit sehr guter Näherung die bei beliebigen Tongemischen auftretende Periodentonhöhe voraussagen, falls die Hüllkurvenperiode des Signals und die Eigenschaften der Tonhöhenempfindung des untersten Teiltons bekannt sind.

Notes: Summary The behaviour of neurones during signal transmission is determined partly by the input to the nerve cell and partly by its inherent properties. A detailed knowledge of the spike trains, which are input to and output from a synapse, may yield information about these synaptic mechanisms. This problem has been treated in relation to data from the dorsal spino-cerebellar tract (DSCT) cells monosynaptically activated from primary endings of muscle spindles. After a description of the firing pattern of these cells the behaviour of various models are examined by computer simulation. A particular type of model approximates the behaviour of the DSCT neurones closely with a rather narrow set of parameters. The model predicts that about 15 primary afferent fibres from one muscle converge on one DSCT cell and that the average size of their EPSPs may be as large as 50% of the threshold of firing.

Notes: Summary 1. Fluctuations in the resting membrane potential of frog nerve fibers were analysed in the frequency range from 1 to 10 000 radians per second. The power spectrum follows a 1/f law. 2. Latency fluctuations have been measured for different stimulus intensities. The relation between standard deviation and mean follows a quadratic law. It can be shown for a number of different receptors that this relation is in part linear in part quadratic.

Notes: Summary The Hodgkin-Huxley theory of ion fluxes across membranes during excitation is extended to explain the graduated depolarisation (receptor potential) of sensory cell membranes. Electric circuit equivalents of living membranes are developed. Driving forces and velocity coefficients are represented by means of electric parameters. From this model active and passive ionic fluxes can be calculated quantitatively on the basis of transport equations derived from irreversible thermodynamics. Thus the circuit equivalent may be used as an analog computer. Electric receptor models allow a reproduction of all potential curves which have been derived in electrophysiological experiments on PD-receptors. The results obtained by the use of this model agree with the results obtained in biological experiments under various conditions of stimuli. The significance of solution compartments for intra-and extracellular ions in relation to the time functions of various conditions are discussed in detail. This models are of heuristic value in experimental research. In combination with neuron networks they can be used for the analysis of information theoretical problems.

Notes: Abstract Clear evidence of electrophysiological effects causing simultaneous contrast phenomena in the visual system of animals has lately been presented. These findings incite to subjective perceptual psychophysiological experiments and give some directions concerning the interpretation and generalizations of the results. Psychophysical measurements have been made of the visual response to sinusoidal, spatially varying stimuli in the mesopic region well above contrast threshold. The results are presented with the characteristic properties, object contrast, spatial frequency, and average luminance of the physical luminance distribution as parameters. Some regular features of the fundamental input-output relations of the visual system are elucidated. Non-linear as well as quasi-linear processes are predicted to be prevalent.

Notes: Summary Information content and information capacity of chromatograms are investigated using the mole fraction information as defined in previous communications. The terminus chromatographic information is introduced in a way that enables the calculation of the total separation work supplied to the mixture. The amount of information is a measure of the effectivity of chromatographs and chromatographic methods under different conditions.

Notes: Summary Some records, obtained from the surface of the optic tectum of the frog with moving visual stimuli are presented as evidence of a global oscillation of the tectal activity whose time course is specific for different patterns of stimulation.

Notes: Summary The transfer properties of the optical system in the arthropod compound eye are determined by the interommatidial angle Δ ϕ, influencing the resolving power, and by the width of the visual fields of single ommatidia Δ ϱ, influencing the response at high spatial frequencies of brightness distributions in the object space. The energy transfer/ receptor is proportional to (Δϕ Δ ϱ)2 and decreases with in-inreasing approximation of the perfect-imaging condition: gDϕ → 0; Δϱ → 0. However, a value Δϕ Δϱ 〉 0 has to be maintained in order to overcome the threshold of nervous excitation at a certain minimum-brightness level. Theoretical treatment yields Δϱ/Δϕ=0.62 to 0.88 as the corresponding optimum-imaging relation. The actual ratio can be derived from measurements of the optomotor reactions to the movement of periodic brightness patterns. The approximate value 0.76 is obtained from the fruitfly Drosophila with normal and mutant eye pigmentation. As a result, the parameters of this imaging system are found to be established in a way that enables optimum performance at sufficient illumination. An dieser Stelle möchte ich Dr. W. Reichardt für sein eingehendes Interesse und manche anregende Diskussion über die Sehvorgänge im Komplexauge meinen Dank sagen. Dr. K. Kirschfeld verdanke ich ebenfalls wertvolle Hinweise. Herrn E. Freiberg bin ich für die Anfertigung der Abbildungen sehr verbunden.

Notes: Summary Recent electrophysiological studies of the activity in response to acoustic stimuli of single primary afferent neurons in the VIIIth nerve of mammals strongly suggest that the spike activity is inherently stochastic in that a) the detailed pattern of activity varies unpredictably from repetition to repetition of an identical acoustic wave-form, but b) appropriate averages of this activity show stability or statistical regularity. Since the only way in which auditory information can reach the more central parts of the nervous system is via the VIIIth nerve, the effective “neural noise” implied by such stochastic “coding” of auditory information must set some sort of limits on auditory discriminations. As shown in this paper, these limits can be calculated if an appropriate statistical model of the neural activity is accepted. Of course, no more than a bound on discrimination performance can be determined in this way, since there is no reason to suppose that the central processing in the nervous system is effectively ideal — particularly in the artificial situations of auditory psychophysics. But if these bounds closely approximate performance (and there is some reason to believe they may, at least for the very simple auditory discriminations treated in this paper) then it seems reasonable to maintain that the observed stochastic “coding” in a sense accounts for or explains these behavioral limits. The key to such an analytical effort as proposed here is the availability of sufficient data on the activity in single primary afferent fibers to permit one to infer a reasonable statistical model describing the total (spike) activity in all the tens of thousands of acoustic fibers of the VIIIth nerve in response to the signals of interest. In spite of extensive studies of single-fiber activity by a number of observers, highfrequency tone bursts (several kilocycles or more) are one of the few classes of signals for which an even roughly adequate amount of data exists at present. For such signals the data suggest that the firing pattern of each neuron can be modeled as a sample function from a simple stationary random process whose parameters depend on stimulus frequency, tensity, cochlear location, neuron sensitivity, etc. The desired overall model can then be obtained by assuming reasonable distributions for these parameters over the population of neurons. Given such a model for the pattern of activity in the auditory nerve, discriminations between tone bursts of different frequencies or different intensities can be studied. These differences will be reflected in the properties of the random processes characterizing the auditory nerve. Using the theory of statistical hypothesis testing, it is possible to determine the smallest difference which could be reliably discriminated by an ideal computer examining the total activity of all fibers. As will be shown, the performance of this ideal computer parallels in a number of ways the performance of human observers. The general trends of this performance, moreover, (including an effect similar to Weber's Law) are largely insensitive to the details of the random process model for the neuron activity. However, it is interesting that the functioning of the ideal computer is particularly and uniquely simple for precisely that class of random process models which seems best to represent the physiological data.

Notes: Summary Information Theory is applied to chemical reactions involving selection processes. It turns out that the information entropy S I is closely related to the corresponding entropy function S T as used in chemical thermodynamics. The change of the molecular information entropy for 1 mol of uniformly synthesized substance is given in „moldit” where Δ S I=Δ ST/R. With n decision in the course of synthesis of a single molecule and z possibilities for each decision there is Δ S I =n lg z. (E.g. The number of possibilities z is 2 if one special optical isomer has to be selected out of the two existing kinds with any decision.) The entropy change is calculated for various selective processes as, for instance, the replication of polymers with a certain lenght in the case of uniform subunits or with a certain composition on sequence; in the case of different subunits. The replication processes require additional selection work that can be derived from the calculated entropy changes.

Notes: Summary Ricciardi et al. [3] have obtained the output distribution functions of a nonlinear switching element with a Poissonian sequence of impulses at the input. They have studied certain limiting properties of these distribution functions. However, certain limiting properties of these distributions become more obvious if one studies the underlying stochastic processes of these distributions. Further, the methods used by Ricciardi et al. for the derivation of the distribution functions and their limiting properties are quite involved. It might perhaps be useful to derive the main results of their paper by alternative simpler methods and point out the underlying stochastic processes more clearly.

Notes: Summary A mathematical model is presented that is supposed to describe certain types of multimodal interval distributions of neuronal discharges. Basically it consits of the selective interaction between an excitatory and an inhibitory impulse sequence. The theoretical results are compared with nerve cell interval distributions reported in the literature, and with distributions from simulation studies. A possible relationship between the properties of this model and longtailed interval distributions is indicated. Several extensions are discussed.

Notes: Summary Repetitive stimulation of human peripheral nerves in situ produces an amplitude oscillation of the evoked action potentials. The purpose of this experiment was to describe the dynamics of the response as a function of the characteristics of the electrical stimuli. Two attempts to define precisely the origin of the response have proved ineffective.

Notes: Abstract Measurement of isometric neck torque of the locust, in response to small sinusoidal motion of visual test patterns with large stripes, shows that displacements of 20 seconds of arc are perceived by the eye. On the other hand, when stripe size is varied, the eye seems not to resolve much detail since no response is elicited by patterns with spatial period less than 3°. It is shown that these two results are not incompatible. Current procedures for comparing geometrical interference phenomena in visual reflexes with the receptor spatial sampling relevant to motion perception are extended to treat the small-signal locust experiment, and shown in general to involve larger confidence limits than usually supposed. Especially, arbitrarily weighted contributions from several ommatidial pair-types in the hexagonal lattice are permissible. Finally, consideration of the effects of receptor and other series nonlinearities on motion-perception experiments of this kind predicts particular test patterns for which visual responses should depend upon phase relations of superposed Fourier spatial components, whether the events of receptor interaction involve correlation or not. Measured effects on the reflex of pattern luminance, contrast, displacement and spatial period form a basis for the small-signal frequency analysis described in the paper which follows this one.

Notes: Zusammenfassung 1. Bestimmte Input-Output Beziehungen wurden untersucht durch intraneurale Registrierungen von isolierten visceralen Ganglien von Aplysia californica. Alle Zellen wiesen eine verlängerte Folge von einem einzigen Typ von EPSPs auf, und lösten infolgedessen aufeinanderfolgende Spikes aus. EPSPs traten entweder (i) spontan auf, oder wurden (ii) ausgelöst durch Erregung eines Konnektifs mittels elektrischer Impulse eines Erregers, der durch einen Geigerzähler betrieben wurde. Die präspike Epoche wurde untersucht, und die Vorgänge, welche den postsynaptischen Spike auslösen, wurden probabilistisch identifiziert. Versuche mit durch digitalen Rechner simulierte Neuronen reproduzierten und erweiterten die Ergebnisse von Tierexperimenten. 2. Die Spike-Wahrscheinlichkeit spiegelt das Verhalten von EPSPs wieder, welche innerhalb eines begrenzten Zeitraumes (als Integrationsperiode bezeichnet) vorkommen und erfaßt daher nur eine begrenzte Anzahl dieser Potentiale (als beeinflussende EPSPs bezeichnet). Die Wahrscheinlichkeit, daß ein gegebenes EPSP einen Spike auslöst, ist im allgemeinen eine abnehmende Funktion des mittleren zeitlichen Abstands (eine zunehmende Funktion der Durchschnittsrate) einer bestimmten Anzahl von kurz zuvor erfolgten EPSPs. Bei gegebener mittlerer Zeitspanne (Durchschnittsrate) wird sie im allgemeinen größer bei Mustern, in denen sukzessive Intervalle immer kürzer werden. Darüber hinaus wird die Spike-Wahrscheinlichkeit beeinflußt durch kurz zuvor erfolgte postsynaptische Spikes; die Wirksamkeit irgendeiner EPSP-Konfiguration kann durch geeignete Anordnung in bezug auf die vorangegangenen Spikes verbessert werden. Es wird daraus gefolgert, daß eine postsynaptische Zelle die Entscheidung, einen Spike zu erzeugen, unter dem Einfluß einer großen Zahl von EPSPs durch laufende Auswertung der genauen Zeitfolge kurz zuvor erfolgter Input-Ereignisse fällt. Deshalb hängt die Bildung der postsynaptischen Spike-Kette von verschiedenen statistischen Besonderheiten der präsynaptischen Entladungen ab. Die Begrenzung dieser Begriffsbestimmung und ihre Anwendbarkeit auf verschiedene, z.T. komplexere Fälle wird diskutiert. 3. Die Arbeitsweise einer synaptischen Verbindung dieses Typs wird gemessen und diskutiert unter besonderer Berücksichtigung der folgenden drei Wahrscheinlichkeitswerte und deren Beziehungen untereinander: (i) die generative Wahrscheinlichkeit, daß eine bestimmte Input-Zeitfolge von präsynaptischen Spikes (oder EPSPs) erfolgt; (ii) die prospektive Wahrscheinlichkeit, daß ein Output-postsynaptischer Spike durch die vorhergehende Input-Zeitfolge ausgelöst wird; (iii) die retrospektive Wahrscheinlichkeit, daß eine bestimmte Input-Zeitfolge erfolgt ist, wenn ein Output-Spike ausgelöst wurde. Jeder dieser Wahrscheinlichkeitswerte (s. Appendix) kann direkt aus den Versuchsergebnissen abgeschätzt werden (a) und hat eine bestimmte physiologische Bedeutung in bezug auf die Eigenschaften der präsynaptischen Neurone, der synaptischen Verbindung, und/oder der postsynaptischen Neurone (b). Die in den prospektiven und retrospektiven Schemata bestehenden Unbestimmtheiten werden gemessen (s. Appendix). Die gefundenen Werte zeigten in welchem Ausmaß die Unbestimmtheit bezüglich Input (Output) reduziert werden kann durch das Bekanntsein von Output (Input) und wie sie verringert werden kann indem die betreffenden Wahrscheinlichkeiten als Funktion der Zeitfolge ausgedrückt werden.

Notes: Summary A mathematical analysis is given of systems of decision equations (such as typically arise in the study of switching networks), with particular regard to the study of self-sustaining periodic solutions, or reverberations, which are of special importance for the description, or simulation, of thought-processes in brains or analogues (“neuronic equations” of ref. 1 in the text). A special class of “normal” equations is singled out for detailed study; these are shown, by means of suitable transformations, to share, notwithstanding their utter non-linearity, the tensorial-product property of linear systems. A criterion is given which permits to design on inspection a network which will admit an arbitrarily pre-assigned reverberation; a condition is also given, which is sufficient for a system to admit of no reverberations having an arbitrarily pre-assigned period.

Notes: Zusammenfassung Das femorale Chordotonalorgan adaptiert nicht vollständig. — Werden die femoralen Chordotonalorgane einer Körperseite operativ dauernd gespannt, weichen die Tiere auf einer senkrechten Lauffläche in Richtung der intakten Körperseite von der Senkrechten ab. — Der Regelkreis zur Stabilisierung des Femur-Tibia-Gelenkes kann auf unterschiedliche Werte adaptieren. Daraus werden Rückschlüsse auf die einzelnen Glieder des Regelkreises gezogen. — Eine Operationstechnik zur Verlegung des Ansatzes der Receptorsehne des femoralen Chordotonalorganes von der dorsalen auf die ventrale Seite des Femur-Tibia-Gelenkes wird beschrieben. Auf diese Weise operierte Tiere bewegen die Femur-Tibia-Gelenke während des Laufens nur, wenn die Tarsen einen Gegenstand berühren. Sonst bleiben die Beine starr ausgestreckt. Sind die Tiere dagegen in Ruhe, beugen sie das operierte Bein in regelmäßigen Abständen. — Die Untersuchung der Tiere am Laufrad ergibt: Verhindert man das Rückschwingen eines Beines, wird es beim Laufen nicht bewegt. Es erzeugt aber eine Kraft. Diese Kraft ist im intakten Bein größer als bei durchtrennter Receptorsehne und kleiner als bei dauernd gespanntem Chordotonalorgan. — Es wird ein Modell entwickelt, das die Steuerung des Führungsgrößengebers für die Bewegung eines Beines durch die bis jetzt bekannten Afferenzen aus Beinreceptoren abbildet.

Notes: Summary On the basis of a recent physical theory of many-body problems developped in our Institute, a model of the brain is formulated, and it is shown how some of its typical features, such as learning and memory processes, find therein a natural and simple explanation. In the Appendix a short surview of the necessary mathematical formalism is finally given.

Notes: Summary The panting mechanism is considered as resulting from the joint effort of the two systems regulating the body temperature and the blood gas tensions respectively. Equations are derived which describe the equilibrium conditions for each system. A nomogram for the evaluation of the amount of heat taken up by one liter of respiratory air is given. Combination of the equilibrium equations leads to a infinite series due to the fact that heat dissipation by the respiratory tract involves increased heat production by the respiratory muscles. The conditions of convergence for the infinite series are derived assuming a quadratic relation between heat production of the respiratory muscles and respiratory minute volume. It is shown that the system will become unstable if the series diverges. Equations for the partial washout of the dead space are given which are essential for the independent control of alveolar ventilation and dead space ventilation by proper adjustment of tidal volume and respiratory rate. Two examples demonstrate the limited value of the panting mechanism as compared with the heat dissipation by sweat production, when the animals are subjected to high environmental temperatures. Panting seems superior however for eliminating an increased heat production due to muscular exercise at very low temperatures as for instance in sled dogs.

Notes: Summary The function of the facet-separating pigments in the compound eyes of the fruitfly Drosophila melanogaster with hypernormal (se), normal (+), subnormal (wa), and missing (w) pigmentation was studied by investigation of: (1) the in-flight optomotor responses to movement of striped patterns with a mean brightness of 300 cd/m2, and (2) the retinal action potentials evoked by flashes in a program of .0003 cd/m2 average brightness. The pigment deficient mutants (w a, w) are less sensitive to the pattern contrast in the bright adapted state, and more sensitive to the flash intensity in the dark adapted state than either the wild-type (+) or the overpigmented mutant(se). These differences are complementary and can be explained by the increased translucency of the pigment cells. Thus the photoreceptors in the equally illuminated eyes of the normal and mutant animals +, se, w a, and w are expected to receive light in a ratio of about 1∶1∶7∶19. However the sensitivity of the receptors as well as the half-peak widths and the density of their visual fields are apparently independent of the eye pigmentation and seem to be equal at common levels of adaptation. The effects of omnidirectional excess light reaching the receptors of the pigment deficient mutants can be simulated in less translucent eyes: when certain amounts of background illumination were combined with the optomotor stimulus in the visual fields of the wild-type receptors it was possible to elicit the predicted “mutant behavior”.

Notes: Zusammenfassung Zur Erfassung der dynamischen Vorgänge bei der Muskelkontraktion wurde das System des fibrillären Insektenflugmuskels unter Bedingungen, die eine Linearisierung zulassen, untersucht. Um die Frequenzantwort als das Systemverhalten im sinusförmigen stationären Zustand zu erhalten, wurden die durch aufgeprägte sinusförmige Längenänderungen erzeugten Spannungsänderungen gemessen. Die Frequenzganganalysen beziehen sich auf die Annahme eines zeitinvarianten Systems mit konzentrierten Elementen. Die dominanten passiven Strukturen des Muskels, die Verbindungs-filamente und die Myosinbrücken, können in dem Frequenzbereich, der den Arbeitsfrequenzen des Insektenflugmuskels entspricht, durch ein aus drei Elementen bestehendes viscoelastisches System des Maxwell-Typs mit Zeitkonstanten vergleichbarer Größe beschrieben werden. Für die Frequenzantwort des aktivierten Muskels wurde mit hinreichender Genauigkeit die Übertragungsfunktion eines Phasenminimum-systems ermittelt. Eine theoretische Übertragungsfunktion für aktive Seite der Kontraktion ist durch Differenzbildung experimentell erhaltener Ortskurven bestätigt worden. Auf der Grundlage dieser theoretischen Differenzkurve wurde eine Beziehung zwischen oscillatorischer Arbeit und enzymatischer Hydrolyse des Adenosintriphosphates (ATP) als Energiequelle hergestellt. Bis zur optimalen Frequenz der Oscillation unter den benutzten Bedingungen wird eine Proportionalität zwischen diesen Größen festgestellt, die durch einen konstanten mechanochemischen Kopplungsfaktor bestimmt ist.

Notes: Abstract A systems theory describing signal transmission in neuronal layer structures is presented. Such structures having the quality of linearity and homogenity can easily be treated by using a multiple Fourier transformation method with respect to the space coordinates and to the time. The characteristic laws of this transformation are discussed. In analogy to linear networks the transfer function, the impulse response function and the step response function are defined which characterizes the layer system. Appropriate test functions are derived. Typical structures of standing and moving patterns are finally discussed. Biological application and simulation of the theory by a system using coherent optics will be presented later.