Humans can rapidly convert instructions about a rule into functional neural structures used to apply the rule. The early stages of this encoding process are poorly understood. We designed a stimulus-response (SR) task in which participants were first shown a SR rule on a screen for 200 ms, and then had to apply it to a test stimulus T, which either matched the S in the rule (SR trial) or not (catch trial). To investigate the early stages of rule encoding, the delay between the end of rule display and the onset of the test stimulus was manipulated and chosen between values of 50 ms to 1300 ms. Participants conducted three sessions of 288 trials each, separated by a median of 9 h. Random sequences of 20 rules were used. We then analysed the reaction times and the types of errors made by participants in the different conditions. The analysis of practice effects in session 1 suggests that the neural networks that process SR and catch trials are at least partially distinct, and improve separately during the practice of respectively SR and catch trials. The rule-encoding process, however, is common to both tasks and improves with the number of trials, irrespective of the trial type. Rule encoding shows interesting dynamic properties that last for 500 ms after the end of the stimulus presentation. The encoding process increases the response time in a non-stochastic way, simply adding a reaction time cost to all responses. The rule-retrieval system is functional before the encoding has stabilized, as early as 50 ms after the end of SR rule presentation, with low response errors. It is sensitive to masking however, producing errors with brief (100 ms) test stimulus presentations. Once encoding has stabilized, the sensitivity to masking disappears. It is suggested that participants do encode rules as a parametrized function, using the same neural encoding structure for each trial, rather than reconfiguring their brain anew for each new SR rule. This structure would have been implemented from instructions received prior to the experiment, by using a library of neural functions available in the brain. The observed errors are consistent with this view.

Electromyogram (EMG)-based Pattern Recognition (PR) systems for upper-limb prosthesis control provide promising ways to enable an intuitive control of the prostheses with multiple degrees of freedom and fast reaction times. However, the lack of robustness of the PR systems may limit their usability. In this paper, a novel adaptive time windowing framework is proposed to enhance the performance of the PR systems by focusing on their windowing and classification steps. The proposed framework estimates the output probabilities of each class and outputs a movement only if a decision with a probability above a certain threshold is achieved. Otherwise (i.e., all probability values are below the threshold), the window size of the EMG signal increases. We demonstrate our framework utilizing EMG datasets collected from nine transradial amputees who performed nine movement classes with Time Domain Power Spectral Descriptors (TD-PSD), Wavelet and Time Domain (TD) feature extraction (FE) methods and a Linear Discriminant Analysis (LDA) classifier. Nonetheless, the concept can be applied to other types of features and classifiers. In addition, the proposed framework is validated with different movement and EMG channel combinations. The results indicate that the proposed framework works well with different FE methods and movement/channel combinations with classification error rates of approximately 13% with TD-PSD FE. Thus, we expect our proposed framework to be a straightforward, yet important, step towards the improvement of the control methods for upper-limb prostheses.

"Dira" is a novel experimental paradigm to record combinations of behavioral and metacognitive measures for the creative process. This task allows assessing chronological and chronometric aspects of the creative process directly and without a detour through creative products or proxy phenomena. In a study with 124 participants we show that (a) people spend more time attending to selected vs. rejected potential solutions, (b) there is a clear connection between behavioral patterns and self-reported measures, (c) the reported intensity of Eureka experiences is a function of interaction time with potential solutions, and (d) experiences of emerging solutions can happen immediately after engaging with a problem, before participants explore all potential solutions. The conducted study exemplifies how "Dira" can be used as an instrument to narrow down the moment when solutions emerge. We conclude that the "Dira" experiment is paving the way to study the process, as opposed to the product, of creative problem solving.

This project aims at the self-orientation of a robot in zero gravity and similar environments towards a given object, stored in its memory as reference object, using an optical camera. The robot has been programmed to self-orient until the object is focused in the centre of the image captured. The positional error of the object in the image captured to its position in a reference image is used to drive the attitude controlling system. This system comprises of a circular disc of a suitable mass attached to a motor to generate a reactive torque causing the robot to rotate in the opposite direction. The final tests carried out in water demonstrated the self-guided orientation by the robot towards the reference object; however, further work needs to be carried out to bring about accuracy and precision in its functioning.

Teleoperated mobile manipulators are of use for disabled people and for the wider public interested in acting at distance. The high price of existing devices is a barrier to their diffusion. The paper reports on the first design produced in the Cheap Arm Project (CHAP). It costs less than £2000, uses easily available parts and can be assembled by anybody with basic technical skills. The manipulator can reach objects from floor-level up to shelves at a height of 170 cm using a new low-cost arm design. Teleoperation is be done using a tablet, smartphone or browser. The cost could be further reduced by using different servo motors. The design and assembly instructions are made available on the open-source repository GitHub, with the hope that the community will build and improve the design. The first version has been tested in a college for disabled young people who provided initial recommendations for improvement.

This paper examines the case of a bipedal robot under an external impact along the axis of the two supporting feet. The dynamics of the robot is modelled using the 3-Mass Linear Inverted Pendulum Model. The model shows that, for impacts below a given threshold, the robot recovers naturally and no corrective action is required. For larger, destabilising impacts, this paper described how to calculate a single or a sequence of corrective steps. The key information used for the calculations is the initial velocity generated by the impact. The behaviour of the model for various initial configurations and impact parameters is illustrated by simulations.

We investigate the problem of achieving robust control of hand prostheses by the electromyogram (EMG) of transradial amputees in the presence of variable force levels, as these variations can have a substantial impact on the robustness of the control of the prostheses. We also propose a novel set of features that aim at reducing the impact of force level variations on the prosthesis controlled by amputees. These features characterize the EMG activity by means of the orientation between a set of spectral moments descriptors extracted from the EMG signal and a nonlinearly mapped version of it. At the same time, our feature extraction method processes the EMG signals directly from the time-domain to reduce computational cost. The performance of the proposed features is tested on EMG data collected from nine transradial amputees performing six classes of movements each with three force levels. Our results indicate that the proposed features can achieve significant reductions in classification error rates in comparison to other well-known feature extraction methods, achieving improvements of ≈ 6% to 8% in the average classification performance across all subjects and force levels, when training with all forces.

Background: People with hereditary and spontaneous spastic paraparesis (HSSP) report that their legs are stiffer and walking is slower when their legs are cold.Objectives: This study explored the effects of prolonged superficial cooling and warming of the lower leg on walking speed and local measures of neuromuscular impairments.Methods: This was a randomised pre- and post-intervention study of 22 HSSP participants and 19 matched healthy controls. On 2 separate occasions, one lower leg was cooled or warmed. Measurements included walking speed and measures of lower limb impairment: ankle movement, passive muscle stiffness, spasticity (stretch reflex size), amplitude and rate of force generation in dorsi-and plantarflexors and central and peripheral nerve conduction time/velocity.Results: For both participants and controls, cooling decreased walking speed, especially for HSSP participants. For both groups, cooling decreased the dorsiflexor rate and amplitude of force generation and peripheral nerve conduction velocity and increased spasticity. Warming increased dorsiflexor rate of force generation and nerve conduction velocity and decreased spasticity.Conclusions: Superficial cooling significantly reduced walking speed for people with HSSP. Temperature changes were associated with changes in neuromuscular impairments for both people with spastic paraparesis and controls. This study does not support the use of localised cooling in rehabilitation for people with spastic paraparesis as reported in other neurological conditions. Rehabilitation interventions that help prevent heat loss (insulation) or improve limb temperature via passive or active means, particularly when the legs and/or environment are cool, may benefit people with spastic paraparesis. (C) 2016 Elsevier Masson SAS. All rights reserved.

This chapter describes the design and evaluation of a learning agent capable of receiving multimodal instructions from a human teaching it how to play a card game. The system design includes a corpus-based grammar, a common representation for both input modalities that is used by the agent's higher-level knowledge, and dialogue components. The initial deployment of the system was as a computer-based agent interacting with human instructors via speech and a touch screen. Our evaluation of this configuration consisted of two experiments involving human users who interacted with the agent using free-flowing speech and card moves executed via the touch screen. The system's recognition of non-verbal input and the pairing of nonverbal input with spoken instructions were virtually flawless. Most problems were out-of-grammar errors. We discuss options for reducing those errors and improving other aspects of the system's operation.

When a humanoid robot walks with pace, different joints face different requirements of speed and torque. This paper shows how to estimate the requirements for three heavily loaded joints: one in the knee and two in the hips. The dynamics of the walking robot is modeled as that of a 3-mass linear inverted pendulum (3-mass LIPM). The 3-mass LIPM is the simplest model that requires considering effects of internal forces and torques dues to swinging masses. The calculations show that the three joints test the characteristics of standard servo motors in different ways, and their requirements can be individually modulated by changing the gait pattern and the robot design.

An algorithm for fitting multiple models that characterize the projective relationships between point-matches in pairs of (or single) images is proposed herein. Specifically, the problem of estimating multiple algebraic varieties that relate the projections of 3 dimensional (3D) points in one or more views is predominantly turned into a problem of inference over a Markov random field (MRF) using labels that include outliers and a set of candidate models estimated from subsets of the point matches. Thus, not only the MRF can trivially incorporate the errors of fit in singleton factors, but the sheer benefit of this approach is the ability to consider the interactions between data points.The proposed method (CSAMMFIT) refines the outlier posterior over the course of consecutive inference sweeps, until the process settles at a local minimum. The inference “engine” employed is a Markov Chain Monte Carlo (MCMC) method which samples new labels from clusters of data points. The advantage of this technique pertains to the fact that cluster formation can be manipulated to favor common label assignments between points related to each other by image based criteria. Moreover, although CSAMMFIT uses a Potts-like pairwise factor, the inference algorithm allows for arbitrary prior formulations, thereby accommodating the needs for more elaborate feature based constraints.

Biological systems are able to recognise temporal sequences of stimuli or compute in the temporal domain. In this paper we are exploring whether a biophysical model of a pyramidal neuron can detect and learn systematic time delays between the spikes from different input neurons. In particular, we investigate whether it is possible to reinforce pairs of synapses separated by a dendritic propagation time delay corresponding to the arrival time difference of two spikes from two different input neurons. We examine two subthreshold learning approaches where the first relies on the backpropagation of EPSPs (excitatory postsynaptic potentials) and the second on the backpropagation of a somatic action potential, whose production is supported by a learning-enabling background current. The first approach does not provide a learning signal that sufficiently differentiates between synapses at different locations, while in the second approach, somatic spikes do not provide a reliable signal distinguishing arrival time differences of the order of the dendritic propagation time. It appears that the firing of pyramidal neurons shows little sensitivity to heterosynaptic spike arrival time differences of several milliseconds. This neuron is therefore unlikely to be able to learn to detect such differences.

The problem of orientation parameterization for applications in computer vision and robotics is examined in detail herein. The necessary intuition and formulas are provided for direct practical use in any existing algorithm that seeks to minimize a cost function in an iterative fashion. Two distinct schemes of parameterization are analyzed: The first scheme concerns the traditional axis-angle approach, while the second employs stereographic projection from unit quaternion sphere to the 3D real projective space. Performance measurements are taken and a comparison is made between the two approaches. Results suggests that there exist several benefits in the use of stereographic projection that include rational expressions in the rotation matrix derivatives, improved accuracy, robustness to random starting points and accelerated convergence.

Around the world, there are thousands of people who lost a hand during war or as a consequence of an accident. Artificial hand prosthesis controlled by surface electromyography (EMG) signals offers a promising solution to improve the quality of life of amputees. As part of the process of prosthesis fitting, an occupational therapist will try to train the amputee with the help of a physical prosthesis [1] that is not actually fitted, but only displayed to provide visual feedback, but these are expensive (>£16,000). This training should be performed for long periods of time at the rehabilitation centre prior to the prosthesis fitting. It aims at improving the generation of nerve signals for capture by EMG probes, and at tuning of the EMG pattern recognition algorithm to the actions most suited for each amputee [2]. This part of the rehabilitation process can be made more efficient and more widely available through the use of a low-cost actuated hand with the same degrees of freedom as the prosthetic device to be fitted.

During bipedal gait, a robot falls from one foot to the other. This motion can be approximated with that of an inverted pendulum with discrete movements of the contact point. We detail here how to use the linear inverted pendulum model (LIPM) for selecting successive contact points in such a way that the centre of mass (COM) of the robot flexibly follows a predefined set of waypoints on a straight or curved trajectory, allowing it to move forward, stop and revert its direction of motion in stable way. The use of a fixed step cycle duration reduces the mathematical complexity and the computational load, enabling real-time updating of gait parameters in a microcontroller.

Although there have been many advances in electromyography (EMG) signal processing and pattern recognition (PR) for the control of multi-functional upper-limb prostheses, some the outstanding problems need to be solved before practical PR-based prostheses can be put into service. Some of these are the lack of training and deployment protocols and the provision of the tools required. Therefore, we present a preliminary procedure to personalize the prosthesis deployment. In the first step, we record the demographic information of each individual amputee person and their background. In the second step of the protocol, the EMG signals are acquired. PR algorithms and parameters will be chosen in the 3(rd) step of the protocol. In the 4(th) step, the best number of EMG sensors to achieve the maximal performance with a full set of gestures is identified. The final step involves finding the best set of movements that the amputee person can produce with an accuracy > 95% as well as identifying the movements with the worst performance, which would require further training. This proposed approach is validated with 2 transradial amputees.

This paper intends to present and analyze all the necessary steps one needs to take in order to build a simple application for structure from motion in practice using only the epipolar constraint. In other words, toillustratein practice (i.e,, cook-book style) about how the most fundamental of concepts in epipolar geometry can be combined into a complete and disambiguated series of steps for camera motion estimation and 3D reconstruction in a relatively simple computer program.In particular, the essential matrix is re-introduced in order to avoid misconceptions regarding the nature of the rotation matrix and translation between two camera positions during consequent computations. Strictly based on the formulation of the essential matrix in this paper, the most important of its properties are given with proofs, merely to provide necessary intuition into the series of derivations that ultimately lead to the recovery of the orientation and baseline vector (up-to-scale). Finally, an application that performs 3D position tracking of a simple web-camera is implemented (and source code made available) in order to demonstrate the practical value of the essential matrix with only a few of assumptions loosely in place (still background and relatively reliable point tracking).

Humans can learn associations between visual stimuli and motor responses from just a single instruction. This is known to be a fast process, but how fast is it? To answer this question, we asked participants to learn a briefly presented (200 ms) stimulus-response rule, which they then had to rapidly apply after a variable delay of between 50 and 1300 ms. Participants showed a longer response time with increased variability for short delays. The error rate was low and did not vary with the delay, showing that participants were able to encode the rule correctly in less than 250 ms. This time is close to the fastest synaptic learning speed deemed possible by diffusive influx of AMPA receptors. Learning continued at a slower pace in the delay period and was fully completed in average 900 ms after rule presentation onset, when response latencies dropped to levels consistent with basic reaction times. A neural model was proposed that explains the reduction of response times and of their variability with the delay by (i) a random synaptic learning process that generates weights of average values increasing with the learning time, followed by (ii) random crossing of the firing threshold by a leaky integrate-and-fire neuron model, and (iii) assuming that the behavioural response is initiated when all neurons in a pool of m neurons have fired their first spike after input onset. Values of m=2 or 3 were consistent with the experimental data. The proposed model is the simplest solution consistent with neurophysiological knowledge. Additional experiments are suggested to test the hypothesis underlying the model and also to explore forgetting effects for which there were indications for the longer delay conditions. This article is part of a Special Issue entitled Neural Coding 2012.

This paper describes findings from a Human-to-Human Interaction experiment that examines human communicative non-verbal facial behaviour. The aim was to develop a more comfortable and effective model of social human-robot communication. Analysis of the data revealed a strong co-occurrence between human blink production and non-verbal communicative behaviours of own speech instigation and completion, interlocutor speech instigation, looking at/away from the interlocutor, facial expression instigation and completion, and mental communicative state changes. Seventy-one percent of the total 2007 analysed blinks co-occurred with these behaviours within a time window of +/- 375 ms, well beyond their chance co-occurrence probability of 23%. Thus between 48% and 71% of blinks are directly related to human communicative behaviour and are not simply "physiological" (e.g., for cleaning/humidifying the eye). Female participants are found to blink twice as often as male participants, in the same communicative scenario, and have a longer average blink duration. These results provide the basis for the implementation of a blink generation system as part of a social cognitive robot for human-robot interaction.

A method for the classification of finger movements for dexterous control of prosthetic hands is proposed. Previous research was mainly devoted to identify hand movements as these actions generate strong electromyography (EMG) signals recorded from the forearm. In contrast, in this paper, we assess the use of multichannel surface electromyography (sEMG) to classify individual and combined finger movements for dexterous prosthetic control. sEMG channels were recorded from ten intact-limbed and six below-elbow amputee persons. Offline processing was used to evaluate the classification performance. The results show that high classification accuracies can be achieved with a processing chain consisting of time domain-autoregression feature extraction, orthogonal fuzzy neighborhood discriminant analysis for feature reduction, and linear discriminant analysis for classification. We show that finger and thumb movements can be decoded accurately with high accuracy with latencies as short as 200 ms. Thumb abduction was decoded successfully with high accuracy for six amputee persons for the first time. We also found that subsets of six EMG channels provide accuracy values similar to those computed with the full set of EMG channels (98% accuracy over ten intact-limbed subjects for the classification of 15 classes of different finger movements and 90% accuracy over six amputee persons for the classification of 12 classes of individual finger movements). These accuracy values are higher than previous studies, whereas we typically employed half the number of EMG channels per identified movement.