Input-output Function Research Articles

PCA-based neuro-fuzzy model for system identification of smart structures

This paper proposes an efficient system identification method for modeling nonlinear behavior of civil structures. This method is developed by integrating three different methodologies: principal component analysis (PCA), artificial neural networks, and fuzzy logic theory, hence named PANFIS (PCA-based adaptive neuro-fuzzy inference system). To evaluate this model, a 3-story building equipped with a magnetorheological (MR) damper subjected to a variety of earthquakes is investigated. To train the input-output function of the PANFIS model, an artificial earthquake is generated that contains a variety of characteristics of recorded earthquakes. The trained model is also validated using the1940 El-Centro, Kobe, Northridge, and Hachinohe earthquakes. The adaptive neuro-fuzzy inference system (ANFIS) is used as a baseline. It is demonstrated from the training and validation processes that the proposed PANFIS model is effective in modeling complex behavior of the smart building. It is also shown that the proposed PANFIS produces similar performance with the benchmark ANFIS model with significant reduction of computational loads.

Non-linear Membrane Properties in Entorhinal Cortical Stellate Cells Reduce Modulation of Input-Output Responses by Voltage Fluctuations.

The presence of voltage fluctuations arising from synaptic activity is a critical component in models of gain control, neuronal output gating, and spike rate coding. The degree to which individual neuronal input-output functions are modulated by voltage fluctuations, however, is not well established across different cortical areas. Additionally, the extent and mechanisms of input-output modulation through fluctuations have been explored largely in simplified models of spike generation, and with limited consideration for the role of non-linear and voltage-dependent membrane properties. To address these issues, we studied fluctuation-based modulation of input-output responses in medial entorhinal cortical (MEC) stellate cells of rats, which express strong sub-threshold non-linear membrane properties. Using in vitro recordings, dynamic clamp and modeling, we show that the modulation of input-output responses by random voltage fluctuations in stellate cells is significantly limited. In stellate cells, a voltage-dependent increase in membrane resistance at sub-threshold voltages mediated by Na+ conductance activation limits the ability of fluctuations to elicit spikes. Similarly, in exponential leaky integrate-and-fire models using a shallow voltage-dependence for the exponential term that matches stellate cell membrane properties, a low degree of fluctuation-based modulation of input-output responses can be attained. These results demonstrate that fluctuation-based modulation of input-output responses is not a universal feature of neurons and can be significantly limited by subthreshold voltage-gated conductances.

A multilevel multimodal circuit enhances action selection in Drosophila

Natural events present multiple types of sensory cues, each detected by a specialized sensory modality. Combining information from several modalities is essential for the selection of appropriate actions. Key to understanding multimodal computations is determining the structural patterns of multimodal convergence and how these patterns contribute to behaviour. Modalities could converge early, late or at multiple levels in the sensory processing hierarchy. Here we show that combining mechanosensory and nociceptive cues synergistically enhances the selection of the fastest mode of escape locomotion in Drosophila larvae. In an electron microscopy volume that spans the entire insect nervous system, we reconstructed the multisensory circuit supporting the synergy, spanning multiple levels of the sensory processing hierarchy. The wiring diagram revealed a complex multilevel multimodal convergence architecture. Using behavioural and physiological studies, we identified functionally connected circuit nodes that trigger the fastest locomotor mode, and others that facilitate it, and we provide evidence that multiple levels of multimodal integration contribute to escape mode selection. We propose that the multilevel multimodal convergence architecture may be a general feature of multisensory circuits enabling complex input-output functions and selective tuning to ecologically relevant combinations of cues.

Neural coding of sound envelope in reverberant environments.

Speech reception depends critically on temporal modulations in the amplitude envelope of the speech signal. Reverberation encountered in everyday environments can substantially attenuate these modulations. To assess the effect of reverberation on the neural coding of amplitude envelope, we recorded from single units in the inferior colliculus (IC) of unanesthetized rabbit using sinusoidally amplitude modulated (AM) broadband noise stimuli presented in simulated anechoic and reverberant environments. Although reverberation degraded both rate and temporal coding of AM in IC neurons, in most neurons, the degradation in temporal coding was smaller than the AM attenuation in the stimulus. This compensation could largely be accounted for by the compressive shape of the modulation input-output function (MIOF), which describes the nonlinear transformation of modulation depth from acoustic stimuli into neural responses. Additionally, in a subset of neurons, the temporal coding of AM was better for reverberant stimuli than for anechoic stimuli having the same modulation depth at the ear. Using hybrid anechoic stimuli that selectively possess certain properties of reverberant sounds, we show that this reverberant advantage is not caused by envelope distortion, static interaural decorrelation, or spectral coloration. Overall, our results suggest that the auditory system may possess dual mechanisms that make the coding of amplitude envelope relatively robust in reverberation: one general mechanism operating for all stimuli with small modulation depths, and another mechanism dependent on very specific properties of reverberant stimuli, possibly the periodic fluctuations in interaural correlation at the modulation frequency.

Cell-Specific Activity-Dependent Fractionation of Layer 2/3→5B Excitatory Signaling in Mouse Auditory Cortex

Auditory cortex (AC) layer 5B (L5B) contains both corticocollicular neurons, a type of pyramidal-tract neuron projecting to the inferior colliculus, and corticocallosal neurons, a type of intratelencephalic neuron projecting to contralateral AC. Although it is known that these neuronal types have distinct roles in auditory processing and different response properties to sound, the synaptic and intrinsic mechanisms shaping their input-output functions remain less understood. Here, we recorded in brain slices of mouse AC from retrogradely labeled corticocollicular and neighboring corticocallosal neurons in L5B. Corticocollicular neurons had, on average, lower input resistance, greater hyperpolarization-activated current (Ih), depolarized resting membrane potential, faster action potentials, initial spike doublets, and less spike-frequency adaptation. In paired recordings between single L2/3 and labeled L5B neurons, the probabilities of connection, amplitude, latency, rise time, and decay time constant of the unitary EPSC were not different for L2/3→corticocollicular and L2/3→corticocallosal connections. However, short trains of unitary EPSCs showed no synaptic depression in L2/3→corticocollicular connections, but substantial depression in L2/3→corticocallosal connections. Synaptic potentials in L2/3→corticocollicular connections decayed faster and showed less temporal summation, consistent with increased Ih in corticocollicular neurons, whereas synaptic potentials in L2/3→corticocallosal connections showed more temporal summation. Extracellular L2/3 stimulation at two different rates resulted in spiking in L5B neurons; for corticocallosal neurons the spike rate was frequency dependent, but for corticocollicular neurons it was not. Together, these findings identify cell-specific intrinsic and synaptic mechanisms that divide intracortical synaptic excitation from L2/3 to L5B into two functionally distinct pathways with different input-output functions.

The Relationship Between EABR and Auditory Performance and Speech Intelligibility Outcomes in Pediatric Cochlear Implant Recipients.

The primary purpose of the present study was to investigate the relationship between postimplant electrically evoked auditory brainstem response (EABR) measures (wave V threshold, wave V latency, input-output functions) and auditory performance and speech intelligibility outcomes measured using parental rating scales, such as Categories of Auditory Performance (CAP; Archbold, Lutman, & Marshall, 1995) and the Speech Intelligibility Rating scale (SIR; Allen, Nikolopoulos, Dyar, & O'Donoghue, 2001), respectively. The secondary purpose was to evaluate the relationship between age at implantation and ratings on the CAP and SIR. Forty children with congenital sensorineural hearing loss participated. Preimplant parental ratings on the CAP and SIR were obtained, and all the children underwent cochlear implantation. Intracochlear EABRs were recorded postimplantation. Postimplant parental ratings on the CAP and SIR were obtained. The relationships between EABR parameters and auditory performance and speech intelligibility outcomes were studied. The effect of age at implantation on auditory performance and speech intelligibility outcomes was also investigated. A significant negative correlation was found between EABR wave V thresholds and SIR growth (r = -.415, p = .016). Children with better CAP growths tended to have lower wave V thresholds than those with poorer CAP growths. Age at implantation had an effect on the auditory performance as measured using the CAP. The present study provides evidence for the relationship between wave V thresholds of the intracochlear EABR and auditory performance and speech intelligibility outcomes measured using parental rating scales in pediatric cochlear implant recipients. Data also indicate that early intervention has a positive impact on auditory performance outcomes.

Neuromorphic log-domain silicon synapse circuits obey bernoulli dynamics: a unifying tutorial analysis.

The field of neuromorphic silicon synapse circuits is revisited and a parsimonious mathematical framework able to describe the dynamics of this class of log-domain circuits in the aggregate and in a systematic manner is proposed. Starting from the Bernoulli Cell Formalism (BCF), originally formulated for the modular synthesis and analysis of externally linear, time-invariant logarithmic filters, and by means of the identification of new types of Bernoulli Cell (BC) operators presented here, a generalized formalism (GBCF) is established. The expanded formalism covers two new possible and practical combinations of a MOS transistor (MOST) and a linear capacitor. The corresponding mathematical relations codifying each case are presented and discussed through the tutorial treatment of three well-known transistor-level examples of log-domain neuromorphic silicon synapses. The proposed mathematical tool unifies past analysis approaches of the same circuits under a common theoretical framework. The speed advantage of the proposed mathematical framework as an analysis tool is also demonstrated by a compelling comparative circuit analysis example of high order, where the GBCF and another well-known log-domain circuit analysis method are used for the determination of the input-output transfer function of the high (4th) order topology.

Orientation selectivity in inhibition-dominated networks of spiking neurons: effect of single neuron properties and network dynamics.

The neuronal mechanisms underlying the emergence of orientation selectivity in the primary visual cortex of mammals are still elusive. In rodents, visual neurons show highly selective responses to oriented stimuli, but neighboring neurons do not necessarily have similar preferences. Instead of a smooth map, one observes a salt-and-pepper organization of orientation selectivity. Modeling studies have recently confirmed that balanced random networks are indeed capable of amplifying weakly tuned inputs and generating highly selective output responses, even in absence of feature-selective recurrent connectivity. Here we seek to elucidate the neuronal mechanisms underlying this phenomenon by resorting to networks of integrate-and-fire neurons, which are amenable to analytic treatment. Specifically, in networks of perfect integrate-and-fire neurons, we observe that highly selective and contrast invariant output responses emerge, very similar to networks of leaky integrate-and-fire neurons. We then demonstrate that a theory based on mean firing rates and the detailed network topology predicts the output responses, and explains the mechanisms underlying the suppression of the common-mode, amplification of modulation, and contrast invariance. Increasing inhibition dominance in our networks makes the rectifying nonlinearity more prominent, which in turn adds some distortions to the otherwise essentially linear prediction. An extension of the linear theory can account for all the distortions, enabling us to compute the exact shape of every individual tuning curve in our networks. We show that this simple form of nonlinearity adds two important properties to orientation selectivity in the network, namely sharpening of tuning curves and extra suppression of the modulation. The theory can be further extended to account for the nonlinearity of the leaky model by replacing the rectifier by the appropriate smooth input-output transfer function. These results are robust and do not depend on the state of network dynamics, and hold equally well for mean-driven and fluctuation-driven regimes of activity.

Characterizing the Input-Output Function of the Olfactory-Limbic Pathway in the Guinea Pig.

Nowadays the neuroscientific community is taking more and more advantage of the continuous interaction between engineers and computational neuroscientists in order to develop neuroprostheses aimed at replacing damaged brain areas with artificial devices. To this end, a technological effort is required to develop neural network models which can be fed with the recorded electrophysiological patterns to yield the correct brain stimulation to recover the desired functions. In this paper we present a machine learning approach to derive the input-output function of the olfactory-limbic pathway in the in vitro whole brain of guinea pig, less complex and more controllable than an in vivo system. We first experimentally characterized the neuronal pathway by delivering different sets of electrical stimuli from the lateral olfactory tract (LOT) and by recording the corresponding responses in the lateral entorhinal cortex (l-ERC). As a second step, we used information theory to evaluate how much information output features carry about the input. Finally we used the acquired data to learn the LOT-l-ERC “I/O function,” by means of the kernel regularized least squares method, able to predict l-ERC responses on the basis of LOT stimulation features. Our modeling approach can be further exploited for brain prostheses applications.

Imaging Submillisecond Membrane Potential Changes from Individual Regions of Single Axons, Dendrites and Spines.

A central question in neuronal network analysis is how the interaction between individual neurons produces behavior and behavioral modifications. This task depends critically on how exactly signals are integrated by individual nerve cells functioning as complex operational units. Regional electrical properties of branching neuronal processes which determine the input-output function of any neuron are extraordinarily complex, dynamic, and, in the general case, impossible to predict in the absence of detailed measurements. To obtain such a measurement one would, ideally, like to be able to monitor, at multiple sites, subthreshold events as they travel from the sites of origin (synaptic contacts on distal dendrites) and summate at particular locations to influence action potential initiation. It became possible recently to carry out this type of measurement using high-resolution multisite recording of membrane potential changes with intracellular voltage-sensitive dyes. This chapter reviews the development and foundation of the method of voltage-sensitive dye recording from individual neurons. Presently, this approach allows monitoring membrane potential transients from all parts of the dendritic tree as well as from axon collaterals and individual dendritic spines.

Weighing fusion method for truck scale based on an optimal neural network with derivative constraints and a lagrange multiplier

This paper presents a novel weighing fusion method for a truck scale based on an optimal neural network with partial derivative constraints and a Lagrange multiplier (PD-LMNN). In this proposed method, firstly, the constraints for optimizing neural network (NN) are constructed via the truck scale’s prior knowledge that the partial derivative of the truck scale’s input–output function is positive. Secondly, the neural network’s performance index is created by using the augmented Lagrange function with a multiplier and a penalty factor. Thirdly, the detail algorithm of training the constrained-optimization NN is given. This proposed method can improve the NN’s generalization ability in case of the lacking training samples. The comparative experimental results show that the weighing errors of the truck scale with PD-LMNN are far less than those of DINN (DINN is a method for training a NN only by using the data samples, not the prior knowledge), and the PD-LMNN’s generalization ability is better than that of DINN when the training samples are lacking.

Is off-frequency overshoot caused by adaptation of suppression?

This study is concerned with the mechanism of off-frequency overshoot. Overshoot refers to the phenomenon whereby a brief signal presented at the onset of a masker is easier to detect when the masker is preceded by a “precursor” sound (which is often the same as the masker). Overshoot is most prominent when the masker and precursor have a different frequency than the signal (henceforth referred to as “off-frequency overshoot”). It has been suggested that off-frequency overshoot is based on a similar mechanism as “enhancement,” which refers to the perceptual pop-out of a signal after presentation of a precursor that contains a spectral notch at the signal frequency; both have been proposed to be caused by a reduction in the suppressive masking of the signal as a result of the adaptive effect of the precursor (“adaptation of suppression”). In this study, we measured overshoot, suppression, and adaptation of suppression for a 4-kHz sinusoidal signal and a 4.75-kHz sinusoidal masker and precursor, using the same set of participants. We show that, while the precursor yielded strong overshoot and the masker produced strong suppression, the precursor did not appear to cause any reduction (adaptation) of suppression. Predictions based on an established model of the cochlear input–output function indicate that our failure to obtain any adaptation of suppression is unlikely to represent a false negative outcome. Our results indicate that off-frequency overshoot and enhancement are likely caused by different mechanisms. We argue that overshoot may be due to higher-order perceptual factors such as transient masking or attentional diversion, whereas enhancement may be based on mechanisms similar to those that generate the Zwicker tone.

Genetic Optimization based Adaptive Fuzzy Logic Control of a pH Neutralization Process

Presence of severe nonlinearity in pH neutralization process makes the problem of pH control a difficult and challenging one. Further, optimized control of such highly nonlinear system requires tuning of various parameters with a suitable evolutionary algorithm (EA). This paper presents genetic algorithm (GA) based unconstrained, continuous and single-objective optimization of conventional and adaptive nonlinear fuzzy logic control (FLC) of pH neutralization process for servo and regulatory operations. The GA optimized adaptive FLC scheme changes the universe of discourse of input-output membership functions based on the region of operation and amount of step change in set-point and load. The performance of controllers are evaluated using integral of squared errors (ISE) and tested for random step changes in set-point and load.

Modulation of the input-output function by GABAA receptor-mediated currents in rat oculomotor nucleus motoneurons.

The neuronal input–output function depends on recruitment threshold and gain of the firing frequency–current (f–I) relationship. These two parameters are positively correlated in ocular motoneurons (MNs) recorded in alert preparation and inhibitory inputs could contribute to this correlation. Phasic inhibition mediated by γ-amino butyric acid (GABA) occurs when a high concentration of GABA at the synaptic cleft activates postsynaptic GABAA receptors, allowing neuronal information transfer. In some neuronal populations, low concentrations of GABA activate non-synaptic GABAA receptors and generate a tonic inhibition, which modulates cell excitability. This study determined how ambient GABA concentrations modulate the input–output relationship of rat oculomotor nucleus MNs. Superfusion of brain slices with GABA (100 μm) produced a GABAA receptor-mediated current that reduced the input resistance, increased the recruitment threshold and shifted the f–I relationship rightward without any change in gain. These modifications did not depend on MN size. In absence of exogenous GABA, gabazine (20 μm; antagonist of GABAA receptors) abolished spontaneous inhibitory postsynaptic currents and revealed a tonic current in MNs. Gabazine increased input resistance and decreased recruitment threshold mainly in larger MNs. The f–I relationship shifted to the left, without any change in gain. Gabazine effects were chiefly due to MN tonic inhibition because tonic current amplitude was five-fold greater than phasic. This study demonstrates a tonic inhibition in ocular MNs that modulates cell excitability depending on cell size. We suggest that GABAA tonic inhibition acting concurrently with glutamate receptors activation could reproduce the positive covariation between threshold and gain reported in alert preparation.

Is age-related hearing loss predominantly metabolic?

Studies in animals have shown that age-related hearing loss is predominantly metabolic in origin. In humans, direct access to the cochlea is not usually possible and so non-invasive methods of assessing cochlear mechanical function are required. This study used a non-invasive assay of cochlear mechanical function, otoacoustic emissions, to examine a metabolic versus hair-cell-loss origin for age-related hearing loss. Three subject groups were examined: adult females with clinically normal hearing, adult females with age-related hearing loss, and adult males with noise-induced hearing loss. Contrasting otoacoustic emission input-output functions were obtained for the three groups, suggesting a causal relationship between age-related hearing loss and strial dysfunction.

Differential effects of static and dynamic inputs on neuronal excitability.

The intrinsic excitability of neurons is known to be dynamically regulated by activity-dependent plasticity and homeostatic mechanisms. Such processes are commonly analyzed in the context of input-output functions that describe how neurons fire in response to constant levels of current. However, it is not well understood how changes of excitability as observed under static inputs translate to the function of the same neurons in their natural synaptic environment. Here we performed a computational study and hybrid experiments on rat bed nucleus of stria terminalis neurons to compare the two scenarios. The inward rectifying Kir current (IKir) and the hyperpolarization-activated cation current (Ih) were found to be considerably more effective in regulating the firing under synaptic inputs than under static stimuli. This prediction was experimentally confirmed by dynamic-clamp insertion of a synthetic inwardly rectifying Kir current into the biological neurons. At the same time, ionic currents that activate with depolarization were more effective regulating the firing under static inputs. When two intrinsic currents are concurrently altered such as those under homeostatic regulation, the effects in firing responses under static vs. dynamic inputs can be even more contrasting. Our results show that plastic or homeostatic changes of intrinsic membrane currents can shape the current step responses of neurons and their firing under synaptic inputs in a differential manner.

Effect of chronic stress on short and long-term plasticity in dentate gyrus; Study of recovery and adaptation

Stress dramatically affects synaptic plasticity of the hippocampus, disrupts paired-pulse facilitation and impairs long-term potentiation (LTP). This study was performed to find the effects of chronic restraint stress and recovery period on excitability, paired-pulse response, LTP and to find probable adaptation to very long stress in the dentate gyrus. Thirty-eight male Wistar rats were randomly divided into four groups of Control, Rest–Stress (21days stress), Stress–Rest (recovery) and Stress–Stress (42days stress: adaptation). Chronic restraint stress was applied 6-h/day. Input–output functions, paired-pulse responses and LTP were recorded from the dentate gyrus while stimulating the perforant pathway. We found that chronic stress attenuated the responsiveness, paired-pulse response and LTP in the dentate gyrus. A 21-day recovery period, after the stress, improved all the three responses toward normal, indicating reversibility of these stress-related hippocampal changes. There was no significant adaptation to very long stress, probably due to severity of stress.

Dynamic fidelity control to the central auditory system: synergistic glycine/GABAergic inhibition in the cochlear nucleus.

GABA and glycine are the major inhibitory transmitters that attune neuronal activity in the CNS of mammals. The respective transmitters are mostly spatially separated, that is, synaptic inhibition in the forebrain areas is mediated by GABA, whereas glycine is predominantly used in the brainstem. Accordingly, inhibition in auditory brainstem circuits is largely mediated by glycine, but there are few auditory synapses using both transmitters in maturity. Little is known about physiological advantages of such a two-transmitter inhibitory mechanism. We explored the benefit of engaging both glycine and GABA with inhibition at the endbulb of Held-spherical bushy cell synapse in the auditory brainstem of juvenile Mongolian gerbils. This model synapse enables selective in vivo activation of excitatory and inhibitory neuronal inputs through systemic sound stimulation and precise analysis of the input (endbulb of Held) output (spherical bushy cell) function. The combination of in vivo and slice electrophysiology revealed that the dynamic AP inhibition in spherical bushy cells closely matches the inhibitory conductance profile determined by the glycine-R and GABAA-R. The slow and potent glycinergic component dominates the inhibitory conductance, thereby primarily accounting for its high-pass filter properties. GABAergic transmission enhances the inhibitory strength and shapes its duration in an activity-dependent manner, thus increasing the inhibitory potency to suppress the excitation through the endbulb of Held. Finally, in silico modeling provides a strong link between in vivo and slice data by simulating the interactions between the endbulb- and the synergistic glycine-GABA-conductances during in vivo-like spontaneous and sound evoked activities.

An Experimental Objective Method to Determine Maximum Output and Dynamic Range of an Active Bone Conduction Implant

Recently, a new active bone conduction implant, the Bonebridge, was introduced. This transcutaneous device is proposed as an alternative to previous percutaneous systems. The current study aims to determine the maximum output (MO) of the Bonebridge by making use of Bonebridge-generated sound pressure levels in the occluded ear canal of the unaided ear. The test setup consisted of audiometry and input-output measurements. These tests were performed on 3 Bonebridge users with conductive or mixed hearing loss (bone-conduction thresholds, ≤ 45 dB HL) at least 3 months after implantation surgery. All the patients were implanted and were evaluated in the Antwerp University Hospital. The MO of the device was determined by measuring input-output functions with a microphone placed in the occluded contralateral ear canal using the Aurical REM system. During testing, the sound processor was fitted in linear amplification mode and with unlimited output to determine the MO and the input dynamic range of the Bonebridge. This experimental setup intends to evaluate the device in a fitting program without compression. The mean MO of the device was 55 dB HL (SD, 6 dB HL) at 0.5 kHz, 61 dB HL (SD, 18 dB HL), 71 dB HL (SD, 10 dB HL) at 2 kHz, and 60 dB HL (SD, 10 dB HL) at 4 kHz. The mean dynamic range of the Bonebridge was 41 (SD, 5) dB HL, 46 (SD, 10) dB HL, 46 (SD, 5) dB HL, and 37 (SD, 16) dB HL for 0.5, 1, 2, and 4 kHz, respectively. In summary, ear canal measures can effectively be used to assess input-output behavior of the Bonebridge. The present study indicates that the MO of the Bonebridge ranges from 55 to 71 dB HL, depending on frequency. Accepting a minimum dynamic range of 35 dB with the Bonebridge, fitting of the Bonebridge in a linear program is advocated in patients with a sensorineural hearing loss component of up to 30 dB HL.

Highly Efficient Selective Assembly Method of Horizontal Stabilizer based on Metamodeling and Particle Swarm Optimization

AbstractSelective assembly is a cost-effective approach for reducing the overall variation and thus improving the quality of an assembled product. Traditional selective assembly requires quite a large number of finite-element analysis (FEA) runs and consequently is rather time consuming when considering every part’s deviations. The paper presents the highly efficient selective assembly method that assists operators in horizontal stabilizer assembling while considering the parts’ fabrication deviations. Metamodeling is employed to build the input–output functions from FEA results. The hybrid metamodel consisted of four metamodels and is constructed to express the mapping relationship between fabrication deviations and assembly distortions. Cross validation of metamodels is conducted using 27 FEA simulations. The quantitative selective assembly suggestion, that is, the selective assembly table, is obtained using the hybrid metamodels and particle swarm optimization (PSO) algorithm. The optimal assembly of t...