Comparing a template approach and complex bandpass filtering for single-trial analysis of auditory evoked M100.

Two methods for single-trial analysis were compared, an established parametric template approach and a recently proposed non-parametric method based on complex bandpass filtering. The comparison was carried out by means of pseudo-real simulations based on magnetoencephalography measurements of cortical responses to auditory signals. The comparison focused on amplitude and latency estimation of the M100 response. The results show that both methods are well suited for single-trial analysis of the auditory evoked M100. While both methods performed similarly with respect to latency estimation, the non-parametric approach was observed to be more robust for amplitude estimation. The non-parametric approach can thus be recommended as an additional valuable tool for single-trial analysis.     

Short and middle latency vestibular evoked responses to acceleration in man

We have succeeded in recording short and middle latency vestibular evoked responses in human subjects. The head was held rigidly in a special, patented head holder, constructed individually for each subject, which gripped the teeth of the upper jaw. The stimulus consisted of 2/sec steps of angular acceleration impulses produced by a special motor with intensities of about 10,000°/sec² and with a rise time of 1–2 msec. The electrical activity was recorded as the potential difference between special forehead and mastoid electrodes having a large, secure contact area with the skin. The activity was digitally filtered and averaged in 2 separate channels by means of a Microshev 2000 evoked response system. The short latency responses, with peaks at about 3.5 msec (forehead positive), 6.0 msec (forehead negative) and 8.4 msec (forehead positive; bandpass: 200–2000 Hz; average of 1024 trials), had amplitudes of about 0.5 μV. The middle latency responses had peaks at about 8.8 msec (forehead positive), 18.8 msec (forehead negative) and 26.8 msec (forehead positive; 30–300 Hz; N = 128 trials), with larger amplitudes (about 15 μV). These responses were consistently recorded in the same subject at different times and were similar in different normal subjects. Strenuous control experiments were conducted in order to ensure that these responses are not artefacts due to the movement of conducting media (head, electrodes and leads) in the electromagnetic field of the motor and are elicited by activation of normal labyrinths. Among other controls, they were not present in a cadaver, in patients with bilateral absence of nystagmus to caloric stimuli and in conducting volumes the size of the human head. They were also not masked by white noise.     

Single-trial latency variability does not contribute to fast habituation of the long-latency averaged auditory evoked potential in the albino rat

Fas habituation (FH) is defined as a general reduction in long-latency, vertex-recorded, averaged auditory evoked potential (AEP) amplitude that occurs in response to the second of a pair of acoustic stimuli. Our laboratory has been studying FH in a variety of human populations with different paradigms and has interpreted it to be a measure of neural attentional mechanism(s) and/or resource allocation related to the processing of cognitive information. We have also reported an analogous phenomenon in the rat. In the present investigation, we examined the relationship between FH (viz., averaged AEP component amplitude decrement) and the single-trial latency variability of the AEP peaks comprising that component. Specifically, AEPs were obtained to 60 paired-tone stimuli from unanesthetized and restrained albino rats previously implanted with chronic skull electrodes. Using a template-matching algorithm similar to that used by Michalewski et al. (Electroenceph. clin. Neurophysiol., 1986, 65:59–71), the latency variability for each animal was computed for the N1 and P2 peaks of the single-trial AEPs that were used to compose the averaged wave form. Findings indicated that (a) there was no difference in single-trial latency variability for these peaks either within or across tones, and (b) there was no relationship between single-trial latency variability for either the N1 or the P2 peaks and the overall peak-to-peak amplitude (N1-P2) of the averaged wave form in response to the second tone. Thus, FH of the N1-P2 (i.e. Peak 2) amplitude in the rat is not due to an increase in latency variability across tones.     

Response entrainment of movement fibers in the optic tract of crayfish

The response properties of jittery movement fibers (JMF) in the crayfish optic tract reacting to a non-moving temporally patterned light were analyzed. The JMFs usually show no response during the regular flickering of stationary light with a flash duration of less than 50 msec when the stimulus frequency is between 4 and 20 per second; however they do respond when the flickering stops if a certain number of flashes have been given. The response appears about 50 msec after the first missing flash, i.e., the latency of the response after the last flash of the train changed from 100 to 300 msec. Thus, the “off” response at the end of the flicker is entrained to the stimulus repetition interval and locked onto the time of the first missing flash. The response of a sustaining fiber to an identical stimulus has quite different features as illustrated in Fig. 2. Some of the fibers show responses to the beginning part of the flicker but not necessarily to each flash, and habituate after several flashes. When a single flash longer than 250 msec is given, the fiber shows an “off” response with about 50 msec latency, as it does to sustained light. Some fibers show a double burst of “off” discharge to single flashes; the first at 50 msec is followed after 120 msec by the second one. However, when the flash duration is between 250 and 50 msec, a single flash elicits little or no response. The latency of the “off” response is as much as 300 msec for short single flashes less than 50 msec. An “on” response to flashes of light is observed when the inter-stimulus interval is more than 5 sec. The responses to the beginning part of flicker train are not simply locked to the just preceding flash except the “on” response to the very first one, but they can be the long latency responses to the flash before that. This response is modified in latency by the succeeding flashes in flicker trains and becomes entrained to the missing flash. Four types of entrainment are classified on the basis of the change in latency from the missing flash with regard to the number of flashes in a train. In most cases, 10 flashes are sufficient to entrain the response to the first missing flash. Non-resposiveness, i.e., habituation, during a regular flicker, may be due to an active inhibitory process, initiated by each succeeding light pulse. The response to the missing flash, therefore results from a disinhibited modified response to the last flash. Some JMFs continue to respond to the flicker even after a considerable number of flashes but only when the repetition interval is about 120 msec corresponding well to the interval of the double burst “off” discharge, thus the JMF has a resonant frequency of about 8 Hz. The JMFs appear to be acting as an irregularity detector in temporal sequence.     

P3b-like potential of rats recorded in an active discrimination task

We investigated whether the potential corresponding to the human P3b could be recorded on the dura mater over the frontal cortex of the rat in an active discrimination task. Rats were trained to press a bar within 1200 msec after cessation of the target tone (1000 Hz) lasting for 800 msec, and to withhold an overt response to the standard tone (2000 Hz). Rats were given intracranial electrical stimulation to the medial forebrain bundle as a reward only when they correctly responded to the target tone. The stimulus probability of the target tone was manipulated at 3 levels: 30, 50 and 70% of all trials. Large, slow positive deflections with peak latency at about 400–500 msec were elicited to the target tone, which were preceded by a large negative potential with its peak around 180 msec, whereas such a clear deflection was not elicited to the standard tone irrespective of the stimulus probability. This positive slow deflection was very similar in morphology to the human P3, and this P3-like potential was significantly larger when the target tone was less frequent (30%) than in the case of more frequent targets (50 and 70%). These results suggest that a potential highly similar to P3b can be recorded in the rat, and that it may provide a useful model to investigate the neurophysiological basis of the human P3b.     

Clinical application of the P3 component of event-related potentials. I. Normal aging

Normal adult volunteer subjects ranging in age from 18 to 90 years participated in a study in which analogous auditory and visual paradigms, with infrequently occurring target and non-target events, were used to elicit event-related potentials (ERPs) with a prominent P3 component. Of the 135 subjects participating, 66 completed both auditory and visual paradigms. The amplitude and latency of P3 were analyzed using average ERPs, single trials (adaptive filter) and principal components analysis (PCA). Age regressions were calculated using measures derived from average ERPs and single trials. Single trial measures were better than average ERP measures in demonstrating age-related changes in P3 latency. There was a significant increase in P3 latency with age of 1–1.5 msec/year. The range of normal P3 latency for a given age (1 S.E. of the regression = 40 msec for the visual target stimuli) was much larger than obtained by other investigators.The visual paradigm produced higher P3 latency/age correlations than the auditory paradigm (visual target r = 0.52, non-target r = 0.42; auditory target r = 0.32, non-target r = 0.33). Within individuals, the amplitude and latency of P3 generated by auditory and visual stimuli were highly correlated, though the visual paradigm produced larger and later P3s than the auditory paradigm.There is an apparent change in the scalp topography of P3 with age. In young adults, P3s to target stimuli have a markedly parietal distribution. The distribution of P3 becomes more uniformly distributed from Pz to Fz with age. This may be due to changes in overlapping components such as the slow wave (SW) rather than to changes in the amplitude of P3 per se.     

Elimination of electrocardiogram contamination from electromyogram signals: An evaluation of currently used removal techniques.

Trunk electromyographic signals (EMG) are often contaminated with heart muscle electrical activity (ECG) due to the proximity of the collection sites to the heart and the volume conduction characteristics of the ECG through the torso. Few studies have quantified ECG removal techniques relative to an uncontaminated EMG signal (gold standard or criterion measure), or made direct comparisons between different methods for a given set of data. Understanding the impacts of both untreated contaminated EMG and ECG elimination techniques on the amplitude and frequency parameters is vital given the widespread use of EMG. The purpose of this study was to evaluate four groups of current and commonly used techniques for the removal of ECG contamination from EMG signals. ECG recordings at two intensity levels (rest and 50% maximum predicted heart rate) were superimposed on 11 uncontaminated biceps brachii EMG signals (rest, 7 isometric and 3 isoinertial levels). The 23 removal methods used were high pass digital filtering (finite impulse response (FIR) using a Hamming window, and fourth-order Butterworth (BW) filter) at five cutoff frequencies (20, 30, 40, 50, and 60 Hz), template techniques (template subtraction and an amplitude gating template), combinations of the subtraction template and high pass digital filtering, and a frequency subtraction/signal reconstruction method. For muscle activation levels between 10% and 25% of maximum voluntary contraction, the template subtraction and BW filter with a 30 Hz cutoff were the two best methods for maximal ECG removal with minimal EMG distortion. The BW filter with a 30 Hz cutoff provided the optimal balance between ease of implementation, time investment, and performance across all contractions and heart rate levels for the EMG levels evaluated in this study.     

Speed and Accuracy of Visual Motion Discrimination by Rats

Animals must continuously evaluate sensory information to select the preferable among possible actions in a given context, including the option to wait for more information before committing to another course of action. In experimental sensory decision tasks that replicate these features, reaction time distributions can be informative about the implicit rules by which animals determine when to commit and what to do. We measured reaction times of Long-Evans rats discriminating the direction of motion in a coherent random dot motion stimulus, using a self-paced two-alternative forced-choice (2-AFC) reaction time task. Our main findings are: (1) When motion strength was constant across trials, the error trials had shorter reaction times than correct trials; in other words, accuracy increased with response latency. (2) When motion strength was varied in randomly interleaved trials, accuracy increased with motion strength, whereas reaction time decreased. (3) Accuracy increased with reaction time for each motion strength considered separately, and in the interleaved motion strength experiment overall. (4) When stimulus duration was limited, accuracy improved with stimulus duration, whereas reaction time decreased. (5) Accuracy decreased with response latency after stimulus offset. This was the case for each stimulus duration considered separately, and in the interleaved duration experiment overall. We conclude that rats integrate visual evidence over time, but in this task the time of their response is governed more by elapsed time than by a criterion for sufficient evidence.     

A fast and reliable method for simultaneous waveform, amplitude and latency estimation of single-trial EEG/MEG data

The amplitude and latency of single-trial EEG/MEG signals may provide valuable information concerning human brain functioning. In this article we propose a new method to reliably estimate single-trial amplitude and latency of EEG/MEG signals. The advantages of the method are fourfold. First, no a-priori specified template function is required. Second, the method allows for multiple signals that may vary independently in amplitude and/or latency. Third, the method is less sensitive to noise as it models data with a parsimonious set of basis functions. Finally, the method is very fast since it is based on an iterative linear least squares algorithm. A simulation study shows that the method yields reliable estimates under different levels of latency variation and signal-to-noise ratio?s. Furthermore, it shows that the existence of multiple signals can be correctly determined. An application to empirical data from a choice reaction time study indicates that the method describes these data accurately.     

Decision Criterion Dynamics in Animals Performing an Auditory Detection Task

Classical signal detection theory attributes bias in perceptual decisions to a threshold criterion, against which sensory excitation is compared. The optimal criterion setting depends on the signal level, which may vary over time, and about which the subject is naïve. Consequently, the subject must optimise its threshold by responding appropriately to feedback. Here a series of experiments was conducted, and a computational model applied, to determine how the decision bias of the ferret in an auditory signal detection task tracks changes in the stimulus level. The time scales of criterion dynamics were investigated by means of a yes-no signal-in-noise detection task, in which trials were grouped into blocks that alternately contained easy- and hard-to-detect signals. The responses of the ferrets implied both long- and short-term criterion dynamics. The animals exhibited a bias in favour of responding “yes” during blocks of harder trials, and vice versa. Moreover, the outcome of each single trial had a strong influence on the decision at the next trial. We demonstrate that the single-trial and block-level changes in bias are a manifestation of the same criterion update policy by fitting a model, in which the criterion is shifted by fixed amounts according to the outcome of the previous trial and decays strongly towards a resting value. The apparent block-level stabilisation of bias arises as the probabilities of outcomes and shifts on single trials mutually interact to establish equilibrium. To gain an intuition into how stable criterion distributions arise from specific parameter sets we develop a Markov model which accounts for the dynamic effects of criterion shifts. Our approach provides a framework for investigating the dynamics of decisions at different timescales in other species (e.g., humans) and in other psychological domains (e.g., vision, memory).     

High-pass filter settings affect the detectability of MLRs in humans

Auditory middle latency responses (MLRs) have been recorded in 217 patients ranging in age from 6 days to 20 years. The probability of obtaining MLR components Na and Pa was higher with a high-pass filter setting of 15 Hz, 12 dB/octave as compared to 3 Hz, 6 dB/octave. This effect was found at all ages tested. Age-related latency effects were apparent with 3 Hz but not 15 Hz filtering.     

Single-trial ERP evidence for the three-stage scheme of facial expression processing

Using a rapid serial visual presentation paradigm, we previously showed that the average amplitudes of six event-related potential (ERP) components were affected by different categories of emotional faces. In the current study, we investigated the six discriminating components on a single-trial level to clarify whether the amplitude difference between experimental conditions results from a difference in the real variability of single-trial amplitudes or from latency jitter across trials. It is found that there were consistent amplitude differences in the single-trial P1, N170, VPP, N3, and P3 components, demonstrating that a substantial proportion of the average amplitude differences can be explained by the pure variability in amplitudes on a single-trial basis between experimental conditions. These single-trial results verified the three-stage scheme of facial expression processing beyond multitrial ERP averaging, and showed the three processing stages of "fear popup", "emotional/unemotional discrimination", and "complete separation" based on the single-trial ERP dynamics.     

Cerebral potentials evoked by rectal distention in humans

The afferent pathways from the rectum can be stimulated and studied after mechanical distention of the rectum. We studied the rectum-brain axis in 24 healthy young adults. The rectum was stimulated with a rectal balloon using inflation volumes of 10 ml and 30 ml air at a stimulation frequency of 0.167 Hz. Additional studies were carried out with 20 ml distention volumes, random stimulation frequency, and stimulation rates of 0.08 Hz and 0.017 Hz. We found two different cortical EPs. An early onset EP was present in 21 of the 24 subjects. P1 latency shortened significantly with increasing distention volumes, but peak amplitudes did not change significantly with larger distention volumes. A late onset EP was present in all subjects with NI latency of 210 ± 15 msec, PI latency of 316 ± 24 msec, and NII latency of 444 ± 34 msec. The different EPs could be due to stimulation of two different visceral afferent pathways which are present in the same individual, due to stimulation of two different fiber populations or due to simultaneous stimulation of afferents in surrounding structures. EP recording after rectal stimulation might be useful in future studies of patients with abnormal rectal sensation.     

P300-like potentials in the normal monkey using classical conditioning and an auditory ‘oddball’ paradigm

Endogenous components of evoked potentials resembling P300 in humans were sequentially studied in 3 cynomolgus monkeys (Macaca fascicularis) using an auditory ‘oddball’ paradigm. The two different auditory stimuli were 500 Hz and 4000 Hz tones, designated as the ‘frequent’ and ‘rare’ stimuli, respectively. The probability of ‘rare’ tone presentation was initially 0.2. We further used probabilities of 0.1, 0.3 and 0.5. The ‘rare’ stimulus was reinforced by electrical stimulation, which followed the onset of the high tone by 700 msec. After 3–5 training sessions, a late positive wave was observed following the ‘rare’ tone. The latency of this P300-like signal was 314±16.2 msec, and teh amplitude 23.6±3.14 μV. The amplitude of this potential was modified by changes in stimulus presentation probability and by withholding reinforcement.     

Latency of the P3 event-related potential: Normative aspects and within-subject variability

A growing body of research has focused on the P3 (P300) event-related potential as an electrophysiological correlate of selective attention. The present investigation examines the intrasubject test-retest reliability of the auditory evoked P3 latency measure for neurologically and audiologically normal young adults aged 22–34 years across test sessions separated by 2–4 weeks (N = 9) and across trials within one test session (N = 20). In a target-selection (‘oddball’) paradigm, subjects mentally counted infrequent 2 kHz tone bursts (targets) randomly interspersed within a sequence of 1 kHz tone bursts (non-targets). A strong positive correlation was demonstrated between latencies of test sessions I and II. An analysis of variance did not demonstrate statistically significant latency differences as a function of either tests of trials for the test-retest group (N = 9). Although analysis of variance demonstrated a statistically significant difference between trials within one test session for 20 subjects, the small mean latency difference between trials (4.7 msec) is interpreted as being clinically unimportant. The stability of P3 latency found in this study over a period of 2–4 weeks supports its application to the study of normal and disordered cognitive processes.     

Moderate prenatal carbon monoxide exposure produces persistent, and apparently permanent, memory deficits in rats

The effects of moderate (150 +/- 2 ppm) prenatal carbon monoxide (CO) exposure (maternal HbCO concentrations of 15.6 +/- 1.1%) on learning and memory were assessed in young and aged adult rats using a two-way active avoidance paradigm. In experiment 1, the prenatal CO-exposed rats at 120 days of age acquired a conditioned avoidance response equally well as control animals in a 100-trial session. However, following a 24-hr interval the CO-exposed rats failed to demonstrate significant retention of the task as indicated by the absence of significant improvement in performance over the indicated by the absence of significant improvement in performance over the previous day; control subjects did show significant retention. In experiment 2, in which 120-day-old animals received 50 training trials per day until a criterion of ten consecutive avoidance responses was met, the prenatal CO-exposed subjects again acquired the task as well as control animals. When tested for retention 28 days later, a significant memory impairment was again observed in terms of trials required to reattain the avoidance criterion as well as in total percent avoidance responding. In neither experiment did an analysis of initial or average latency to escape the footshock stimulus reveal any significant alterations. These latter results suggest that the observed performance impairment reflected a memory deficit and not a disruption of sensory, motor, or motivational factors. In experiment 3, prenatal CO-exposed rats approximately 1 year of age (300-360 days of age) showed impairment relative to air-exposed controls in both the original learning and retention of the two-way avoidance response. Again, however, there was no evidence for alterations in performance factors per se. Collectively these data indicate that while young adult rats prenatally exposed to 150 ppm CO demonstrate an associative deficit restricted to memory impairment, aged adults similarly exposed during the prenatal period display a more pronounced deficit similar to that recently reported for animals tested as juveniles. The importance of parametric manipulations in uncovering long-term toxicity is also discussed.     

Questions regarding the sequential neural generator theory of the somatosensory evoked potential raised by digital filtering

Digital bandpass filtering (300–2500 Hz) designed for zero phase shift was applied to somatosensory evoked potentials recorded with cephalic bipolar montages. Four consistent negative and corresponding positive peaks with latencies of about 16, 18, 19, and 20 msec were elicited with median nerve stimulation. Peroneal nerve stimulation also elicited 4 reproducible negative-positive peaks having latencies of about 24, 26, 28, and 30 msec. Interpeak latencies measured 1.3 ± 0.2 msec and 1.8 ± 0.25 msec for median and peroneal elicited SEPs respectively. Becaise cephalic bipolar recordings cancel most far-field potentials, multiple generators cannot account for all the additional components seen. It is hypothesized that some of the high frequency components recorded are due to activity in recurrent intrathalamic neuronal networks.     

An algorithm for defining the onset and cessation of the flexion-relaxation phenomenon in the low back musculature

The flexion-relaxation phenomenon (FRP) in the low back provides insights into the interplay between the active and passive tissues. Establishing a reliable algorithm for defining the lumbar angle at which the muscles deactivate and reactivate was the focus of the current paper. First, the EMG data were processed using six different smoothing techniques (no smoothing, moving average, moving standard deviation, Butterworth low pass filter at 0.5 Hz, 5 Hz, and 50 Hz) herein called the processed EMG (pEMG). The FRP points were then defined using four thresholds (pEMG less than 3% MVC, pEMG less than 5% MVC, pEMG less than 2 times FRP pEMG, and pEMG less than 3 times FRP pEMG). Finally, a duration requirement was tested (no duration requirement, pEMG data must maintain threshold requirement for 50 data points). Each combination of smoothing, threshold, and duration were applied through a computer program to each muscle for all trials and established an EMG-off and EMG-on angle for each muscle. These estimates were compared to the gold standard of expert-identified EMG-off and EMG-on angles and the root mean square error (RMSE) between this gold standard and the predictions of the algorithms served as the dependent variable. The results showed that the most important factor to produce low values of RMSE is to utilize a Butterworth low pass filter of 5 Hz or less and, if this is employed, there is no value to a duration requirement. The results also suggest that using the "3 times FRP pEMG" threshold technique may provide further improvements in these predictions.     

Effects of attention and slow potential shifts on self-regulation of event-related potentials

Research on the effects of self-regulation of slow potentials (SP) and event-related potentials (ERP) has failed to look at the possible interactions of these two kinds of brain potentials. The present study investigated such interactions by recording both ERP and SP potential changes in an operant ERP conditioning paradigm. Ten subjects participated in two conditions that were designed to differentially manipulate attention to the stimuli. In the operant conditioning task, subjects received auditory feedback as they attempted to increase the ERP amplitude at 180 msec poststimulus (P180), which was elicited by a subpainful shock stimulus to the forearm over 250 trials. In the distraction task, subjects were instructed not to attend to stimuli or feedback tones, but rather received and were tested on reading materials. Attention, as manipulated by these tasks, was not a determinant of changes in ERP amplitude since there were no significant differences in the size of P180 between attention conditions. While no significant change in the mean ERP amplitude occurred, subjects were able to produce ERPs above criterion threshold significantly more often during trials in the conditioning task than in the reading task. Thus, there was evidence of some learning. The difference in wave forms between hit and miss trials indicates a latency shift (with misses having a later ERP peak). This may indicate that latency, rather than, or in addition to, amplitude, is shaped during conditioning procedures. In addition, the CNV that developed between the shock stimulus and the feedback signal during conditioning was significantly larger in amplitude than in the distraction condition. This is taken as evidence of increased attention during conditioning. Since hit trials demonstrated larger contingent negative variation (CNV) amplitudes, production of CNVs may be instrumental in mediating hits. Therefore, attentional mechanisms may play a role in successful ERP self-regulation. No correlations were found involving P180, CNVs, or tonic slow potential shifts. Changes in tonic DC levels showed a suggestive trend between conditions. Although both conditions began with a negative shift, during conditioning the negativity increased, while during distraction the tonic level went to positivity. These trends support the hypothesis that attention and arousal increased during conditioning. The possible reasons for the lack of significant correlations between ERP and tonic or phasic slow potential changes in this paradigm are discussed.     

A P300-based Brain-Computer Interface with Stimuli on Moving Objects: Four-Session Single-Trial and Triple-Trial Tests with a Game-Like Task Design

Brain-computer interfaces (BCIs) are tools for controlling computers and other devices without using muscular activity, employing user-controlled variations in signals recorded from the user’s brain. One of the most efficient noninvasive BCIs is based on the P300 wave of the brain’s response to stimuli and is therefore referred to as the P300 BCI. Many modifications of this BCI have been proposed to further improve the BCI’s characteristics or to better adapt the BCI to various applications. However, in the original P300 BCI and in all of its modifications, the spatial positions of stimuli were fixed relative to each other, which can impose constraints on designing applications controlled by this BCI. We designed and tested a P300 BCI with stimuli presented on objects that were freely moving on a screen at a speed of 5.4°/s. Healthy participants practiced a game-like task with this BCI in either single-trial or triple-trial mode within four sessions. At each step, the participants were required to select one of nine moving objects. The mean online accuracy of BCI-based selection was 81% in the triple-trial mode and 65% in the single-trial mode. A relatively high P300 amplitude was observed in response to targets in most participants. Self-rated interest in the task was high and stable over the four sessions (the medians in the 1st/4th sessions were 79/84% and 76/71% in the groups practicing in the single-trial and triple-trial modes, respectively). We conclude that the movement of stimulus positions relative to each other may not prevent the efficient use of the P300 BCI by people controlling their gaze, e.g., in robotic devices and in video games.