Low frequencies improve intensity discrimination for electrocutaneous artificial sensory feedback

The current standard of care for those living with upper-limb loss is unsatisfactory, with up to 50% of amputees abandoning their prostheses, citing unintuitive use and a lack of sensory feedback as critical factors. Electrocutaneous stimulation uses electrodes on the skin to evoke tactile sensation. Stimulation of reinnervated afferent nerves can be used to provide amputees with natural, intuitive somatosensory feedback. Conveying the magnitude of tactile stimuli is an essential characteristic of natural touch. Tactile stimuli that vary in other properties can be judged on a single intensive continuum. Accordingly, the modulation of stimulus frequency results in differences in perceived magnitude of touch. Here, we are exploring the ability to discriminate between stimulation frequencies to restore sensory feedback with sensorized prostheses. Participants received stimulation through a custom-fabricated stimulation pad placed on the palm or residual limb. We measured the just-noticeable difference (JND) to describe how well electrocutaneous stimulation can convey the magnitude of tactile stimuli. The JND is defined as the minimum change in stimulation frequency that can be identified correctly 75% of the time. We quantified the JND using a two-alternative forced-choice paradigm in which stimulation pulse frequency was varied, and participants were asked to determine which of the two pulse frequencies felt stronger. Weber's law states that the JND is proportional to the absolute magnitude of that stimulus and that the percent change relative to the absolute magnitude (known as the Weber fraction) is consistent across stimulus magnitude (i.e., the more intense a stimulus, the greater the change must be to be detected). Here, we show that Weber's law does not iii hold true for at least electrocutaneous stimulation, where the Weber fraction is much smaller at lower stimulus frequencies (10% and 13% change needed at 25 Hz and 50 Hz, respectively, vs 28% change needed at 75 Hz vs 34% change needed at 100 Hz). This suggests that the number of perceivable sensory gradations may be closer to 42 (based on the variable Weber fractions across frequency), which is triple the previously estimated 14 (based on a static Weber fraction of 0.34 obtained at 100 Hz). In contrast, leading invasive methods of stimulation yield an estimated 36 and 15 gradations, respectively. This indicates that low-cost, non-invasive electrocutaneous stimulation constitutes an effective alternative. These results also help deepen the understanding of tactile perception across frequency, where lower frequencies may provide supplemental temporal cues that aid discrimination. These results can help facilitate the implementation of electrocutaneous stimulation as artificial sensory feedback. Improved discrimination of tactile features will benefit neural prostheses in real-world tasks that rely on somatosensory feedback. Its implementation will offer improved patient outcomes for those experiencing upper-limb loss and begin to restore the complete sensory experience of an intact hand. Electrocutaneous stimulation may also constitute a valuable rehabilitation tool for stroke patients and can be used to convey social touch information in telepresence and augmented or virtual reality.

LOW FREQUENCIES IMPROVE INTENSITY DISCRIMINATION FOR ELECTROCUTANEOUS ARTIFICIAL SENSORY FEEDBACK by Abigail R. Citterman A Senior Honors Thesis Submitted to the Faculty of The University of Utah In Partial Fulfillment of the Requirements for the Honors Degree in Bachelor of Science In The Department of Biomedical Engineering Approved: ______________________________ Jacob A. George, PhD Thesis Faculty Supervisor _____________________________ David W. Grainger, PhD Chair, Department of Biomedical Engineering _______________________________ Kelly W. Broadhead, PhD Honors Faculty Advisor _____________________________ Sylvia D. Torti, PhD Dean, Honors College March 2022 Copyright © 2022 All Rights Reserved ABSTRACT The current standard of care for those living with upper-limb loss is unsatisfactory, with up to 50% of amputees abandoning their prostheses, citing unintuitive use and a lack of sensory feedback as critical factors. Electrocutaneous stimulation uses electrodes on the skin to evoke tactile sensation. Stimulation of reinnervated afferent nerves can be used to provide amputees with natural, intuitive somatosensory feedback. Conveying the magnitude of tactile stimuli is an essential characteristic of natural touch. Tactile stimuli that vary in other properties can be judged on a single intensive continuum. Accordingly, the modulation of stimulus frequency results in differences in perceived magnitude of touch. Here, we are exploring the ability to discriminate between stimulation frequencies to restore sensory feedback with sensorized prostheses. Participants received stimulation through a custom-fabricated stimulation pad placed on the palm or residual limb. We measured the just-noticeable difference (JND) to describe how well electrocutaneous stimulation can convey the magnitude of tactile stimuli. The JND is defined as the minimum change in stimulation frequency that can be identified correctly 75% of the time. We quantified the JND using a two-alternative forcedchoice paradigm in which stimulation pulse frequency was varied, and participants were asked to determine which of the two pulse frequencies felt stronger. Weber’s law states that the JND is proportional to the absolute magnitude of that stimulus and that the percent change relative to the absolute magnitude (known as the Weber fraction) is consistent across stimulus magnitude (i.e., the more intense a stimulus, the greater the change must be to be detected). Here, we show that Weber’s law does not ii hold true for at least electrocutaneous stimulation, where the Weber fraction is much smaller at lower stimulus frequencies (10% and 13% change needed at 25 Hz and 50 Hz, respectively, vs 28% change needed at 75 Hz vs 34% change needed at 100 Hz). This suggests that the number of perceivable sensory gradations may be closer to 42 (based on the variable Weber fractions across frequency), which is triple the previously estimated 14 (based on a static Weber fraction of 0.34 obtained at 100 Hz). In contrast, leading invasive methods of stimulation yield an estimated 36 and 15 gradations, respectively. This indicates that low-cost, non-invasive electrocutaneous stimulation constitutes an effective alternative. These results also help deepen the understanding of tactile perception across frequency, where lower frequencies may provide supplemental temporal cues that aid discrimination. These results can help facilitate the implementation of electrocutaneous stimulation as artificial sensory feedback. Improved discrimination of tactile features will benefit neural prostheses in real-world tasks that rely on somatosensory feedback. Its implementation will offer improved patient outcomes for those experiencing upper-limb loss and begin to restore the complete sensory experience of an intact hand. Electrocutaneous stimulation may also constitute a valuable rehabilitation tool for stroke patients and can be used to convey social touch information in telepresence and augmented or virtual reality. iii TABLE OF CONTENTS ABSTRACT ii INTRODUCTION 1 BACKGROUND 3 MATERIALS AND METHODS 13 RESULTS 20 DISCUSSION 27 ACKNOWLEDGMENT 34 REFERENCES 35 iv INTRODUCTION More than one in every 200 individuals suffers from limb loss in the United States alone, with the prevalence projected to triple over the next 30 years [1]. Artificial limbs, or prostheses, are often prescribed to help restore function in amputees. However, nearly 50% of individuals provided with upper-limb prostheses abandon these devices, citing a lack of sensory feedback as a primary reason [1]–[4]. Sensory feedback enhances fine motor skills, reduces phantom limb pain, and fosters a sense of embodiment with one’s prosthesis [5], [6], all of which are necessary for providing fully restorative prosthetic solutions to upperlimb amputees [3]. Artificial sensory feedback offers both functional and psychological benefits to those experiencing upper-limb loss [7]–[10]. Recent research has demonstrated that somatosensory feedback can be effectively conveyed through electrical stimulation with the use of sensorized myoelectric prostheses (i.e., controlled by electromyographic (EMG) signals from musculature in the residual limb) [7]–[15]. The leading method of such stimulation utilizes electrodes implanted in or around peripheral nerves in the residual arm. Implanted electrodes can provide high-resolution, naturalistic sensory feedback, but are expensive and highly invasive [14]–[16]. Since this method requires surgical implantation, there is a risk of complications such as infection [16]. Conversely, electrocutaneous stimulation uses electrodes on the skin to evoke tactile sensation through afferent nerves [14]. This low-cost, non-invasive method of stimulation offers intuitive somatosensory feedback with improved accessibility and minimal risk. Conveying the magnitude of tactile stimuli is an essential characteristic of natural touch. Magnitude discriminability is a critical component of restoring fine motor skills such as those required for fragile object manipulation [7], [8], [10]. Stimulus intensity provides necessary information about the pressure one is exerting on such an object. Previous experiments have explored how stimulation parameters affect the quality and percept of sensory information [14], [15]. However, these studies were limited to invasive methods and did not provide an understanding of the variance in discrimination across a span of intensities or the distinct levels of intensity such stimulation can provide. There remains a gap regarding the optimal electrocutaneous stimulation parameters to most intuitively convey tactile information. To this end, we explored the ability to modulate electrocutaneous stimulation frequency to better convey the magnitude of tactile stimuli to users of sensorized prostheses. We quantified the just-noticeable differences in intensity by measuring human participants’ ability to discriminate between stimulation pulse frequencies. We focused specifically on how that ability changes over a range of stimulus intensity. This information was then used to estimate the maximal number of perceivable sensory gradations that can be conveyed. Together, our findings offer insight as to the optimal stimulation parameters for electrocutaneous stimulation. These parameters can help guide the implementation of intuitive, non-invasive artificial sensory feedback with the use of sensorized prostheses. Such sensory feedback aims to address the many challenges upper-limb amputees face with regards to fine motor skills, phantom limb pain, and embodiment with prosthesis use. This work represents an essential step in restoring the complete sensory experience of an intact human hand for those experiencing upper-limb loss. Electrocutaneous stimulation also has broader impacts for stroke rehabilitation and sensory communication in telepresence and augmented or virtual reality. 2 BACKGROUND Clinical Relevance As of an estimate in 2005, one in 190 individuals in the United States was experiencing limb loss [1]. The prevalence of limb loss in the United States was projected to double by 2020 and more than triple by 2050 in conjunction with a 25% increase in the incidence of vascular disease [1]. However, more recent studies have indicated an increase in vascular disease prevalence closer to 40% [17], which can be attributed to rising levels of obesity, cancer, and other comorbidities [18]–[22]. Further, early findings have shown that COVID-19 infection, and especially hospitalization, significantly increase the risk of vascular disease and related events [23]–[25]. Altogether, these external factors suggest that the prevalence of amputation will likely rise even more drastically in the coming years than projected. The need for adequate prosthetic solutions must increase accordingly. Of those experiencing limb loss, more than one in three are upper-limb amputees [1], [26]. Upper-limb loss is unique in the challenges it presents, both physically and psychologically. One aspect is the complexity and diversity of the functions that are performed by hands [27]. Another is their salience in communication and self-presentation [28]. Consequently, those living with upper-limb loss undergo dramatic changes in i) physical function, ii) self-image, and iii) overall quality of life [27], [29]–[31]. Firstly, physical function is affected by upper-limb loss. Fifty-five distinct validated outcome measures for the functionality of individuals with upper-limb loss exist [32]. Across all metrics, amputation significantly inhibits the performance of activities of daily life (ADLs) [33]–[36]. While prosthesis use is proven to augment functional outcomes 3 [33], many tasks remain a challenge for prosthetic users. Body-powered prostheses, which harness the force and movement of joints proximal to the amputation, are the most prevalent prosthetic solution. These systems can operate mechanical terminal devices [37], but often have limited degrees of freedom (DOFs) and do not restore fine motor skills [38]. Consequently, many functional challenges persist. Secondly, psychological challenges arise in conjunction with self-perception, with prosthetic users citing a lack of unity between themselves and their artificial limbs [5], [27], [28], [39], [40]. It is understood that poor functionality is a critical factor in such feelings of disembodiment [27]. Interestingly, this experience is also observed in total joint replacement and organ transplants, though to a significantly lesser extent [41], [42]. This observed disparity in embodiment implies that artificial parts can be perceived as an extension of one’s body; with a fostered sense of unity, a prosthesis may also be seen as a a part of the body. Thirdly, upper-limb loss impacts overall quality of life. The presence of chronic neuropathic pain is a great challenge for upper-limb amputees. Phantom limb pain is a widely explored but poorly understood phenomenon in which pain persists in the part of the body that has been removed [27], [29], [30], [39], [43]–[45]. Experts believe that the cause of such pain is a combination of peripheral and central neuraxial changes [45], though there is much controversy surrounding the true mechanism [44]. One strong hypothesis points to a loss of input; simply, the deafferentation (interruption or destruction of peripheral sensory nerve fibers that result in a loss of sensory input) causes physiological distress [45]. Though the mechanism responsible for phantom limb pain is likely more complex, this hypothesis is regarded as an acceptable working theory. Chronic phantom 4 pain was reported to occur with between 41% [43] and 80% [44], [45] of upper-limb amputees. Further, limb loss affects overall mental health. The levels of anxiety and depression in amputee populations are significantly higher than the general population [43], [46]. Phantom pain is cited as a contributing factor in the prevalence of such mental illness and the aforementioned sense of disembodiment [46]. Prosthetic solutions must seek to address these persistent challenges. Bodypowered prostheses have not succeeded in solving these issues [37], [47], with rates of upper-limb rejection as high as 45% [2]–[4]. Such abandonment rates serve as a metric of the general amputee population’s satisfaction with their care. Many of these challenges, however, begin to be addressed by myoelectric prostheses. With improved dexterity and control as compared to body-powered terminal devices [12], [48]–[50], mean prosthetic abandonment rates for myoelectric device users are lower than any other population [2], [3]. Nevertheless, there still exists great room, and need, for improvement [51]. Rather than naturalistic muscle activation to control the prosthesis, many bionic hands utilize patterned activation [52] or even foot controls [53]. Without any feedback beyond that of visual, such myoelectric control can be difficult and unintuitive [11]. Recent surveys have examined patients’ desire for sensory feedback [54]–[57], with as many as 88% of upper-limb prosthesis users expressing interest. Implementing sensory feedback is a critical next step in addressing these persistent challenges and providing the most restorative prosthetic solutions. 5 Artificial Sensory Feedback Artificial sensory feedback can further improve functionality and dexterity [13], [58]–[63], foster a greater sense of unity [62], [64], [65], and address phantom limb pain [5], [9], [66], [67] for those experiencing upper-limb loss. Altogether, these factors offer the possibility of complete restoration of an intact hand’s feel and function [68], and thus, enhanced patient rehabilitation. To achieve this, artificial sensory feedback must recreate the tactile experience of the human hand. In an intact individual, mechanoreceptors in the hand are activated by changes in pressure [69]. Merkel cells, Ruffini nerve endings, Pacinian corpuscles, and Meissner corpuscles in the hand are all activated during natural contact with objects [70], receiving stimuli through mechanical transduction [69], [71]. These mechanical forces are rapidly transformed into electrical signals that generate action potentials [72]. The action potentials then propagate along the afferent nerves toward the central nervous system, through the spinal cord, and to the somatosensory cortex in the brain, where the signals are consciously perceived as touch [72] (Fig 1). The greater the force of the stimulus, the greater the number of action potentials triggered, which is perceived as a more intense sensation [69]. Conversely, in an individual experiencing upper-limb loss, the mechanoreceptors in the hand no longer exist. The afferent nerves that transport the signals to the brain, however, remain functional. Artificial sensory feedback bypasses traditional mechanical transduction and applies stimuli directly to the peripheral nerves, providing its own electrical encoding (Fig. 1). The action potentials then propagate along the same natural pathways described above. Embedded sensors in the prosthesis measure the force of the 6 Fig. 1: Sensory pathways for the perception of touch. (A) The perception of touch in an intact individual: (i) initiates with the activation of mechanoreceptors in the hand due to external stimuli; (ii) electrically encodes the mechanical forces; and (iii) propagates as action potentials to the somatosensory cortex of the brain where it is consciously perceived as touch. (B) Conversely, the perception of touch in an individual with upper-limb loss is achieved through artificial sensory feedback, which: (i) begins with embedded sensors in the prosthesis detecting external stimuli; (ii) provides its own electrical encoding and applies stimulation directly to afferent nerves in the residual limb; and (iii) propagates along the same natural pathways to the brain where it is perceived as touch. stimulus and deliver stimulation accordingly. Incorporating sensory feedback allows the bionic hand to convey intuitive information to the user. Notably, this includes the magnitude of tactile stimuli across a span of intensities. Various methods of stimulation have been explored to restore sensory feedback in this capacity, bypassing traditional mechanical transduction and applying stimuli directly to the nerves [16], [73]–[75]. A leading method is invasive stimulation via implanted electrodes [70]. 7 One example of invasive stimulation uses intraneural electrodes. These electrodes are surgically implanted in the nerve, penetrating the nerve’s protective sheath – the epineurium – and directly contacting the afferent fibers [76]. Various interfaces differing in electrode density and layout have been developed to optimize specificity [70]. A leading intraneural device is the Utah slanted electrode array (USEA), which consists of a 10 𝗑𝗑 10 grid of electrodes that penetrate the nerve at different depths (Fig. 2). The USEA offers the most selective activation of nerves for naturalistic feedback [60], [62]. However, such electrodes can be damaging to the nerve fibers due to implantation trauma as well as biocompatibility concerns — the long-term stability of the interface remains in question [77], [78]. An alternative method of invasive stimulation utilizes epineural electrodes. These electrodes are implanted around the nerve and do not penetrate the perineurium. A common type is a cuff electrode, such as the flat interface nerve electrode (FINE) (Fig. 2). FINEs tend to cause minimal nerve damage, encircling and reshaping the fibers as opposed to penetrating them; epineural electrodes also boast more long-term stability [7], [15], [79]. However, due to the way it encompasses, entire nerve bundles are simultaneously activated with these electrodes [80]. This phenomenon results in a lack of selectivity, failing to elicit intuitive sensations [80]. Activation across the afferent population is highly unnatural, often evoking paresthesia [7], colloquially known as “pins and needles.” 8 Fig. 2: Leading invasive methods of stimulation. (A) Intraneural stimulation, illustrated with the Utah slanted electrode array (USEA). The USEA is implanted directly into the residual nerve, penetrating the perineurium. Its 10 𝗑𝗑 10 grid of electrodes boasts selective activation, but the long-term stability remains an issue. (B) Epineural stimulation, illustrated with the flat interface nerve electrode (FINE). The FINE is implanted around the nerve, like a cuff. It offers long-term stability but does not offer selectivity, as the entire fascicle is simultaneously activated. Though implanted electrodes such as USEAs and FINEs can offer high-resolution feedback [10], [13]–[15], [79], [81], their inaccessibility and invasiveness remain challenges. There is a need for a low-cost, non-invasive alternative to provide naturalistic sensory feedback to the general upper-limb amputee population. Such a development is key to helping restore the complete sensory experience of an intact hand. Electrocutaneous Stimulation Non-invasive electrocutaneous stimulation has existed in the clinical and experimental realms since the end of the 19th century, when sound-generating devices (e.g., tuning forks) were used as vibratory generators [82]. During the early 20th century, electromagnetic relays were adapted to produce stimuli [83]. This method was followed by the implementation of concentrated air puffs [83]. Then, in 1927, the early telephone was 9 modified to create a stimulating device, and the first systematic study of electrocutaneous sensitivity was performed [84]. This initial research focused on the potential to “hear” with one’s sense of touch, as motivated by the ability to perceive music through vibrations [85]. The findings, though early, showed great promise and inspired the further exploration of such stimulation as a means of communication. Over the next 40 years, numerous studies were performed on conveying meaningful information through electrocutaneous stimulation [86]–[90]. Despite its tactile nature, the focus was on alternate senses; the applications successfully spanned from a surrogate for braille to the transmission of Morse code [86], [87]. It was not until the 1970s that electrocutaneous stimulation was implemented to provide somatosensory feedback [91]–[95]. This transition from conveying auditory to tactile information marked the beginning of electrocutaneous stimulation’s role in restoring sensation for those experiencing limb loss. These early studies explored modulation of stimulus polarity [91], [93], intensity [89], [94], and electrode count as well as placement [89], [93], [95]. For use with sensorized myoelectric prostheses, these investigations yielded promising functional results in addition to subjective results; participants experiencing limb loss reported great satisfaction with the feedback systems [89], [92], [93]. However, some notable challenges persisted. First, there remained a lack of fundamental understanding of the physiological mechanisms by which tactile sensation was evoked [91]. This ambiguity hindered the thorough exploration of stimulus parameter optimization. Secondly, there was little reproducibility due to inconsistencies in delivered stimuli and in electrode placement [93], which made widespread implementation difficult. 10 Furthermore, these early methods of electrocutaneous stimulation lacked resolution and resulted in a mere three perceivable sensory gradations: “nothing, touch, and deep pressure” [89]. These large, discrete steps in percept were not intuitive, as an intact hand has a much more continuous span of gradations [14], [96]. Finally, several studies reported pain or discomfort affiliated with stimulation [86], [88]–[92], [95]. Such pain indicated the sensations elicited were not natural. With suboptimal parameters, inconsistency, unintuitiveness, and pain; electrocutaneous stimulation was not yet conducive for longterm patient use. Recent studies on electrocutaneous stimulation have worked to address many of these limitations. The physiological mechanisms are better understood today; consequently, stimulation location and parameters can be selected to elicit specific sensations [97]. To that end, several studies have modulated stimulation site, frequency, and current to try to achieve the most consistent, intuitive feedback [98]–[103]. The results between experiments varied, indicating poor reproducibility across stimulation devices and controls. Multichannel stimulation offered higher resolution feedback and improved control but was inconsistent and yielded no change in sensation quality [98], [104]. Other work explored the sensation elicited over time with electrocutaneous stimulation to investigate the viability of long-term implementation with prostheses [105], [106]. Results suggested that discernibility over time depended on stimulation intensity, but that longterm use with prostheses would be possible with the appropriate parameters [105]. To restore intuitive sensory feedback through electrocutaneous stimulation, the optimal stimulation parameters must first be better understood. 11 This work seeks to explore the ability to modulate electrocutaneous stimulation frequency to better convey the magnitude of tactile stimuli to users of sensorized prostheses. We quantified the just-noticeable differences in intensity by measuring individuals’ ability to discriminate between stimulus frequencies, focusing specifically on how that ability changes over the span of intensities. This information was then used to estimate the maximal number of perceivable sensory gradations that can be conveyed. Together, our findings can offer insight as to the optimal stimulation parameters for electrocutaneous stimulation. This can help guide the future implementation of intuitive, non-invasive artificial sensory feedback. Improved discrimination of tactile features will benefit neural prostheses in real-world tasks that rely on somatosensory feedback, an important development in restoring the complete sensory experience of an intact human hand. 12 MATERIALS AND METHODS A. Human Participants Twenty-one intact subjects and one transradial amputee volunteered for this study. The twelve male (55%) and ten female (45%) participants ranged from ages 18 to 60. Additional information regarding the participants and their respective stimulation parameters is located in Table 1. Table 1. Participant Demographics Age Gender BMI (kg/m2) Impedance (kΩ) Stim Threshold (mA) Stim Current (mA) 20 Male 29.8 24.6 1.3 2.8 30 Female 24.1 15.6 1.2 2.7 27 Male 23.1 18.5 1.5 3.0 24 Female 25.9 14.5 1.6 2.9 21 Female 22.3 16.8 1.8 3.3 33 Female 20.7 20.6 1.6 3.1 36 Male 25.3 18.5 1.5 3.0 21 Female 22.5 21.7 1.8 3.3 23 Female 24.2 32.6 1.9 2.9 21 Male 25.2 63.6 3.4 3.9 39 Male 25.8 41.0 3.6 3.9 42 Male 32.5 33.1 2.4 3.9 18 Male 23.5 49.9 2.9 3.4 21 Female 23.7 14.9 1.8 3.3 22 Female 16.6 16.8 1.7 3.2 24 Male 20.9 18.9 1.9 3.4 39 Male 25.1 22.9 2.2 3.7 13 Age Gender BMI (kg/m2) Impedance (kΩ) Stim Threshold (mA) Stim Current (mA) 19 Female 31.9 20.7 1.4 2.9 21 Male 27.3 19.9 2.2 3.7 27 Male 22.5 33.1 1.8 3.3 23 Female 29.4 18.5 1.9 3.4 60 Male 27.2 12.0 6.1 12.2 It has been shown that electrocutaneous stimulation of intact individuals and those with amputations yields similar results in discriminability [14]. Therefore, these studies can be performed on healthy participants and appropriately extrapolate to the amputee population. Informed consent and experimental protocols were carried out in accordance with the University of Utah Institutional Review Board (Protocol No. 110994, Transcutaneous Stimulation for Haptic Feedback). B. Stimulation Device The subjects received electrocutaneous stimulation through a custom-fabricated stimulation pad (Table 2) secured on the palm of the hand or on the residual limb, such that the elicited sensation was perceived as though it was coming from the palm (Fig. 3). The square 9-cm2 silicon stimulation pad consisted of one 0.79-cm2 (1-cm diameter) stimulating electrode surrounded by four 0.44-cm2 (0.75-cm diameter) ground electrodes (Fig. 3). Stimulation was delivered using the Summit Neural Interface Processor with a Micro+Stim front end (Ripple Neuro LLC, Salt Lake City, UT). The signal was amplified prior to delivery with a custom amplifier (modified from [107]). 14 Table 2. Cost Analysis, Stimulation Pad (in USD) Nickel-plated brads (x5) Platinum-cure silicone rubber Flexible wire Heat shrink BNC connector Total* $0.30 $2.00 $1.00 $0.20 $0.50 $4.00 * excluding cost of stimulation device and custom amplifier, as this will vary depending on set-up of choice Fig. 3: Custom-fabricated electrocutaneous stimulation pad. The square 9-cm2 silicon stimulation pad consisted of one 0.79-cm2 stimulating electrode surrounded by four 0.44-cm2 ground electrodes. Placement is illustrated with an intact and an amputee participant, with shaded areas in red depicting the location of perceived sensation. C. Stimulation Parameters Electrocutaneous stimulation was delivered to the central stimulating electrode secured on the base of the palm between the thumb and index finger or forearm (Fig. 3) [14], with caution as to not cause muscle activation. The outer four electrodes served as 15 ground electrodes for return current. The stimulation consisted of biphasic, cathodic-first pulses with 100-µs durations (Fig. 4) [14], [108]. The stimulation pulse frequency varied between 6 and 250 Hz [15]. The stimulation current varied per participant, as it was chosen individually to account for variance in skin impedance. When possible, the current was set 1.5 mA above the threshold of detection (Table 1) [14]. However, at the consequent current, some individuals found the stimulation to be either indiscernible (i.e., too weak) or uncomfortable (i.e., too strong); individual adjustments in stimulation current were made as needed. Fig. 4: Stimulation waveform. Biphasic, cathodic-first pulses with a constant 100-µs phase duration and 100-µs interphase duration were delivered. The current was individually determined for each participant based on individual sensitivity, and the period depended upon the frequency of the pulse delivered as dictated by the experimental paradigm. 16 D. Experimental Design Discriminability was quantified using a two-alternative forced-choice paradigm in which pulse frequency was varied. Two one-second stimulus trains were delivered, separated by an inter-stimulus interval of one second, and participants were asked to determine which of the two one-second stimulus trains felt stronger [15]. There was no time limit for their responses, though they were encouraged to answer promptly to the best of their ability. Participants were also instructed to focus solely on the intensity of the pulse, ignoring any perceived changes in duration or location — tactile stimuli can be judged on a single intensive continuum regardless of differences in modality or quality [109]. The two stimulus trains had constant current and pulse width, varying only in pulse frequency. In each trial, one of the two trains was a fixed reference frequency (25 Hz, 50 Hz, 75 Hz, and 100 Hz) throughout the experiment. The other of the two stimulus trains was a test frequency, ranging from 25% to 175% of the reference frequency. For a given experiment, we explored nine test frequencies (four of a lower intensity than that of the reference, the reference itself, and four greater) as established in [15]. The order of the two stimulus trains in a given trial was randomized, as was the order of the test frequencies. There was a ten-second pause after the response before delivering the next pair of stimuli to minimize the effects of adaptation [106], [110]. A total of 180 trials (20 trials for each of the nine test frequencies) were performed in a single experiment at a given reference frequency. 17 E. Just-Noticeable Difference and Weber Fraction The discrimination data were fit with cumulative normal distributions to psychometric functions. The just-noticeable difference (JND) was estimated as the minimum change in stimulation frequency that could be identified correctly 75% of the time [111], as quantified by the participant responses to the pairs of stimuli. Each psychometric function yielded two JNDs (an upper and a lower JND depicting increases and decreases in pulse frequency), which were averaged. Weber’s law states that the subjective correlate of a just-noticeable difference is constant over the continuum [112]; simply, the JND is proportional to the absolute magnitude of the stimulus. We computed the Weber fraction accordingly in order to compare the discriminability across intensities, independent of the reference frequency. The Weber fraction, defined as the JND divided by the reference frequency (Eq. 1), is given as a percent change that remains consistent across stimulus magnitude. Weber fraction = JND / reference frequency Eq. 1 F. Sensory Gradations To calculate the total number of perceivable gradations, the minimum and maximum stimulation was defined. An iterative, forward-step process from 1 Hz to the maximum stimulation frequency calculated the number of gradations (Algorithm 1). For static Weber fractions across frequency (i.e., Weber's Law holds true and the Weber fraction is consistent), the step size was identified by the JND. For variable Weber fractions (i.e., Weber's law does not hold true and the Weber fractions change with respect to frequency), there were multiple windows, one for each experimental JND. The step size 18 Algorithm 1. Estimating the number of sensory gradations 1: 2: 3: 4: 5: 6: 7: 8: Define min and max frequency Start at min frequency (1 Hz) Multiply by Weber fraction of nearest reference frequency Add to previous frequency Verify result is less than max frequency (300 Hz) Repeat as needed using Weber fraction from nearest reference frequency Stop when result exceeds max frequency (300 Hz) Count number of steps from min to max frequency within each window was identified by the frequencies’ respective JNDs, and the total number of gradations was found using that same iterative process. G. Statistical Analyses All data were screened for normality with the Anderson-Darling test prior to analyses. JNDs at each reference frequency were reported as the mean across all participants (n = 7) ± STD. A one-way ANOVA multiple comparison test (k = 4) with the Dunn-Sidák correction was performed to test for significant differences in mean Weber fractions between all reference frequencies. Due to the small sample size of reported literature Weber fractions for invasive correlates, Grubbs’ test for outliers was used to determine if the intraneural or epineural Weber fractions were statistical outliers relative to our experimental electrocutaneous Weber fractions. 19 RESULTS Low Frequencies Improve Intensity Discrimination We quantified the just-noticeable differences (JNDs) at four reference frequencies to explore the performance across a range of intensities. Systematic changes in the stimulation intensity yielded accordingly systematic changes in the perception of said intensity, as evidenced by the smooth psychometric functions seen below (Fig. 5). The lowest reference frequency, 25 Hz, resulted in the smallest mean JND (2.59 ± 1.21 Hz), exhibiting the sharpest curve. This is followed by 50 Hz (JND = 6.71 ± 2.15 Hz), 75 Hz (JND = 21.33 ± 9.08 Hz), and 100 Hz, the highest reference frequency, which exhibited the largest JND (34.48 ± 13.32 Hz) and, consequently, the least-steep curve (Fig. 5). To compare intensity discriminability across reference frequencies, we calculated the Weber fraction, defined as the JND divided by the reference frequency (Eq. 1). Intensity discrimination is improved at lower frequencies, with smaller changes in frequency able to be reliably detected. 20 Fig. 5: Psychometric functions relating intensity discrimination performance to changes in frequency. Discriminability is given as the probability of the test stimuli being identified as stronger. Pulse frequency (PF) is reported as a percentage relating to the reference frequency. Data indicates the probability of test stimuli being identified as stronger than the reference over the 20 pairwise trials at each of the nine comparisons (100% vs 0%, 90% vs 0%, etc.). Intensity discrimination was enhanced at lower frequencies, as steeper psychometric curves indicate improved performance. Here, we found that Weber’s law does not hold true for electrocutaneous stimulation, where the Weber fraction is significantly smaller at lower stimulus frequencies (10% and 14% change needed at 25 Hz and 50 Hz, respectively, vs 28% change needed at 75 Hz vs 34% change needed at 100 Hz (Fig. 6)). We determined that the sensitivity, as quantified from the Weber fraction, was variable across the range of frequencies with significantly heightened sensitivity at lower frequencies (p < 0.0125). 21 Fig. 6: Intensity discrimination is improved at low frequencies, exhibiting lower Weber fractions. Weber fractions are shown here at each of the four reference frequencies ± standard deviation, with significantly different discriminability between the lower two frequencies and the higher two, *p-value < 0.0125, one-way ANOVA multiple comparison tests; n = 7 for each reference frequency. Electrocutaneous Stimulation Yields Lower Weber Fractions and a Higher Number of Sensory Gradations than Leading Invasive Methods The Weber fractions were compared across the various methods of stimulation to identify the discriminability performance of electrocutaneous stimulation with respect to the leading invasive methods discussed earlier — intraneural, with the Utah slanted electrode array (USEA) [14], and epineural, with the flat interface nerve electrode (FINE) [15], [81]. A 10% change in intensity was required to reliably detect that change at 25 Hz for electrocutaneous stimulation, compared to a 20% change required for epineural stimulation at a similar reference frequency (20 Hz). At the 50 Hz reference frequency, a 22 14% change was required for electrocutaneous stimulation, in contrast to a 10% change for intraneural and a 33% change for epineural. No studies with intraneural or epineural stimulation were performed with a 75 Hz reference frequency, so no comparisons were made with this reference. At the 100 Hz reference frequency, a 34% change was required for electrocutaneous stimulation, as opposed to a 25% change for intraneural and a 30% change for epineural (Fig. 7). Weber fractions were found to be similar for electrocutaneous and intraneural stimulation and for electrocutaneous and epineural stimulation (Grubbs’ test, p > 0.05), with the exception of epineural at 50 Hz, which was found to be an outlier from the electrocutaneous data at that reference frequency (Grubbs’ test, p < 0.05). 23 Fig. 7: Weber fractions for intensity discrimination across methods of stimulation. The discriminability of electrocutaneous stimulation (left) compared to intraneural stimulation via the Utah slanted electrode array (center) and epineural stimulation via the flat interface nerve electrode (right), the leading invasive methods of stimulation. Electrocutaneous data shows mean ± standard deviation. Electrocutaneous stimulation yields Weber fractions similar to intraneural and epineural stimulation (p-value > 0.05, Grubbs’ test for outliers; n = 7 for electrocutaneous bars; n = 1 for intraneural and epineural bars). To explore the maximum number of sensory gradations with electrocutaneous stimulation, we had participants discriminate perceived intensity as a function of pulse frequency. We quantified the JNDs, calculated the Weber fractions, and were then able to calculate the number of perceivable sensory gradations. Our findings suggest that the number of perceivable sensory gradations is closer to 42 (based on the variable Weber 24 fractions across frequency (Table 3)). This is triple the previously estimated 14 perceivable sensory gradations (based on a static Weber fraction of 0.34 obtained at 100 Hz). A higher number of sensory gradations results in more natural and continuous steps, as opposed to a low number of gradations, which relay large and discrete steps. An intact hand would ideally yield a smooth psychophysical function of perceived intensity over the span of external stimuli. Contrastingly, two gradations would appear as a unit step function (Heaviside function), or simply “on/off.” So few gradations would not convey sufficient somatosensory feedback. Electrocutaneous stimulation boasts 42 sensory gradations, as opposed to 36 and 15 achievable gradations with intraneural and epineural stimulation, respectively (Fig. 8). Table 3. Electrocutaneous Weber Fractions and Resultant Sensory Gradations Static Weber Fractions Reference Frequency (Hz) 25 50 75 100 Weber Fraction 0.10 0.14 0.28 0.34 55 38 18 14 Number of Sensory Gradations 25 Variable Fractions Across Frequencies 42 Method of Stimulation Number of Sensory Gradations Electrocutaneous 42 Intraneural 36 Epineural 15 Frequency (Hz) Fig. 8: Electrocutaneous stimulation offers the highest number of sensory gradations as compared to the invasive alternatives. The number of sensory gradations for all three methods of stimulation was estimated from their respective calculated Weber fractions. The y-axis depicts arbitrary units of discriminability, and the log plot x-axis is frequency. Overlaid vertical lines depict the frequencies at which Weber fractions change, i.e., where the step size changes. Electrocutaneous stimulation (red) offers the highest number of gradations, followed by intraneural (light gray) and epineural (dark gray). 26 DISCUSSION Current prosthetic solutions for upper-limb amputees fail to restore the sensory experience of an intact hand, contributing to high prosthesis abandonment rates [1]–[4]. This challenge begins to be addressed through artificial sensory feedback with electrical stimulation [7]–[10]. While the leading method of such stimulation is expensive and highly invasive [14]–[16], low-cost electrocutaneous stimulation uses electrodes on the skin to evoke sensation [14]. This non-invasive alternative offers more accessible sensory feedback. However, the optimal electrocutaneous stimulation parameters are still not wellunderstood [14], [15]. To that end, we explored the ability to modulate electrocutaneous stimulation frequency to better convey the magnitude of tactile stimuli to users of sensorized prostheses. We quantified the just-noticeable differences in intensity by measuring participants’ ability to discriminate between stimulation frequencies. This information was then used to estimate the maximum number of perceivable sensory gradations that can be conveyed. We found that: i) intensity discrimination is enhanced at lower stimulation frequencies; and ii) electrocutaneous stimulation constitutes a low-cost, non-invasive alternative in providing intuitive sensory feedback. These findings can help guide the future implementation of artificial sensory feedback, representing an essential step in restoring a natural sensory experience for individuals experiencing upper-limb loss. We sought to elucidate how the ability to detect changes in intensity varies over the span of external stimuli. This information would provide insight into the optimal stimulation parameters for the most intuitive sensory feedback. Therefore, we collected discrimination data at four reference frequencies (25 Hz, 50 Hz, 75 Hz, and 100 Hz). Lower stimulation frequencies were found to significantly improve discrimination performance 27 (Fig. 5). A 10% and 14% change in frequency was needed at 25 Hz and 50 Hz, respectively, in contrast to a 28% and a 34% change needed at 75 Hz and 100 Hz (Fig. 6). We determined that the sensitivity, as quantified from the Weber fraction (the normalized just-noticeable difference), was variable across the range of frequencies with significantly heightened sensitivity at lower frequencies (p < 0.0125). This demonstrates that Weber’s law does not hold true for electrocutaneous stimulation, as the percent change needed to perceive that change is not constant (Table 3). The improvement of discriminability shown here indicates that delivering stimulation at low frequencies yields more perceptible differences. Being able to detect subtle changes in external stimuli is critical for the successful completion of tasks that rely on somatosensory feedback, such as fragile object manipulation [9], [10], [58], [62] and object recognition [113]. This enhanced sensitivity will augment prothesis users’ fine motor skills and may also address persistent challenges of phantom pain and prosthesis disembodiment [9], [64], [66]–[68]. Previous work in the area of pain and disembodiment focuses on invasive methods, relying on surgically-implanted electrodes [9], [64] or targeted reinnervation [68]. The invasive nature of these methods carries with it a high cost and heightened risk of infection, rendering them inaccessible to the general patient population [14]–[16]. Electrocutaneous stimulation is a low-cost, non-invasive alternative (Table 2). However, in order for electrocutaneous stimulation to constitute an appropriate surrogate for the invasive methods, it must still offer sufficiently high-resolution sensory feedback. To that end, we next sought to examine the resolution of the elicited sensations across the different methods of stimulation. We compared the Weber fractions obtained with non-invasive electrocutaneous stimulation to previous findings for chronically28 implanted intraneural [14] and epineural [15], [81] stimulation, the leading invasive methods (Fig. 2). We found that, at the 50 Hz reference frequency, electrocutaneous stimulation required a 14% percent change in intensity to be able to detect that change, whereas epineural stimulation needed 33% [15] (Fig. 7). This epineural Weber fraction was a statistical outlier from the electrocutaneous data at that reference frequency (Grubbs’ test, p < 0.05), which indicates enhanced discriminability with electrocutaneous stimulation. There were no statistical outliers between electrocutaneous stimulation and either of the invasive methods at the other reference frequencies (Grubbs’ test, p > 0.05), demonstrating similar sensitivity across these methods. From the Weber fractions, we then estimated the maximal number of sensory gradations that could be achieved across all methods of stimulation to quantify the naturalness of the percepts. An intact hand would ideally yield a smooth psychophysical function across the span of external stimuli. However, using reported Weber fractions of force discrimination with an intact hand, there are only an estimated 52 sensory gradations [114], [115]. In contrast, a method of artificial sensory feedback that yielded two gradations would result in discrete “on/off.” Electrocutaneous stimulation yielded 42 gradations, closest to that of an intact hand. In contrast, intraneural resulted in 36 gradations and epineural, only 15 (Fig. 8). While no studies have systematically compared all three methods, recent work has explored the difference in sensitivity between these invasive methods, as well as non-invasive stimulation compared with natural feedback. Intraneural stimulation was shown to outperform epineural in terms of sensitivity [116]. Surprisingly, non-invasive stimulation was found to outweigh natural feedback in dexterity and precision 29 of control [65]. This supports our findings that the implementation of non-invasive stimulation constitutes a naturalistic and effective method of restoring sensory feedback. Selectivity is a major distinction between the three methods presented here. Epineural stimulation activates the entire nerve bundle, eliciting sensation projected to larger areas of skin [7], [80]. This often results in paresthesia, or “pins and needles,” which may mask more subtle percepts. In contrast, intraneural and electrocutaneous stimulation activate fewer afferent nerve fibers, resulting in a smaller projection field [60], [62]. This may allow for the enhanced ability to perceive more subtle changes and, consequently, the elicitation of higher-resolution sensory feedback. While this work represents a critical step in restoring sensory feedback for individuals with upper-limb loss, there remain a few notable limitations. First, the participant population is not entirely representative of the end-user. Though discriminability data from intact individuals was found to appropriately extrapolate to amputees [14], subjective remarks from the intended patient population offer great insight for guiding future development. For example, intact participants would not experience pain with the elicitation of sensation in the same way that an amputee with phantom limb pain or a neuroma may. In this study, we were only able to recruit one amputee participant out of 22 total subjects. Individuals experiencing upper-limb loss must continue to be involved with future studies to ensure patient suitability. Another important limitation is that the manner in which stimulation was delivered may provide supplemental timing cues that aid in discrimination. Participants can detect spaces between the individual pulses at lower stimulation frequencies, perceiving more of a tapping sensation. These spaces can no longer be felt at higher frequencies, eliciting more 30 of a vibratory sensation. Consequently, enhanced discriminability at lower frequencies may in part be attributed specifically to frequency rather than overall intensity [117]. Further studies should seek to isolate frequency discrimination by delivering stimulation with pseudo-random current [81] instead of holding stimulation current constant throughout the experiment. This may offer a more comprehensive understanding of how to best convey the magnitude of tactile stimuli. The present study provides insight as to the optimal electrocutaneous stimulation parameters and highlights the advantages of this non-invasive method of artificial sensory feedback. First, we found that low frequencies improve intensity discrimination. This suggests that stimulating at such frequencies will enhance prosthesis users’ ability to detect changes in the magnitude of stimuli, thus refining their fine motor skills. Such refinement will aid in the completion of real-world tasks that rely on somatosensory feedback. We then used the discrimination data to estimate the maximum number of sensory gradations that could be achieved. We found that electrocutaneous stimulation most closely approaches that of an intact hand, offering the highest number of gradations and the most natural elicitation of sensation of the methods presented here. This demonstrates that lowcost, non-invasive electrocutaneous stimulation constitutes an effective alternative for restoring sensory feedback to individuals experiencing upper-limb loss. Future development should explore the continued advancement of electrocutaneous stimulation using biomimetic patterns, similar to that which was shown in recent invasive studies [10], [70], [118], [119]. Such patterns more closely mimic the natural mechanism of perception and touch [15], [63]. Identifying the optimal patterns may allow for the most intuitive sensory feedback. This may further address the challenges with upper-limb 31 prosthesis use such as hindered dexterity, phantom limb pain, and a sense of disembodiment. Future work should also seek to expand the impact of this promising technology. Development is already underway for an exciting new application of electrocutaneous stimulation: rehabilitation. Stroke is among the top ten leading causes of serious long-term physical and cognitive disability amongst adult Americans [120], comprising nearly 3% of the U.S. population [121]. A prevalent long-term effect is hemiparesis, or partial paralysis of one side of the body, affecting an estimated 80% of stroke survivors [122]. Over 50% of hemiparetic patients present with discriminative somatosensory loss [123], which significantly hinders functional recovery post-stroke [124]. Few quantitative metrics exist to define the extent of sensory loss or track rehabilitation over time. The Rivermead Assessment of Somatosensory Perception is the most widely implemented, but it lacks precision and reliability [125], [126]. Using the same experimental paradigm as presented in this study, we hope to provide a quantitative metric of sensory loss for hemiparetic stroke patients. We will collect discrimination data from both the affected and unaffected hand soon after stroke to quantify the sensory loss with relation to the individual’s baseline. Throughout the patient’s rehabilitation journey, we will repeatedly measure their discrimination performance to track their recovery progress. Electrocutaneous stimulation will serve as a low-cost, non-invasive tool for quantifying sensory loss and subsequent rehabilitation. Such quantification may allow for the validation of future therapies and offer a more comprehensive understanding of stroke rehabilitation over time. Electrocutaneous stimulation also offers a possible pathway for augmenting sensory feedback, which has been shown to be an effective therapy in stroke rehabilitation [127], [128]. 32 Future development may also extend electrocutaneous stimulation to a more mainstream population through applications in telepresence and augmented or virtual reality (AR/VR). Facebook Reality Labs is a leader in the effort to build the future of connection and communication in AR/VR spaces [129]. Reality Labs is exploring various methods of non-invasive stimulation to convey social touch information remotely [130], [131], interact more immersively with AR/VR environments [132], [133], and even enhance the precision of freehand text-entry on a mid-air QWERTY keyboard [134]. Early prototypes within these studies are unintuitive or uncomfortable to use and often utilize costly materials and interfacing components [130]–[134]. Electrocutaneous stimulation may constitute an intuitive and accessible alternative method for such applications. There are significant physical and emotional challenges associated with upper-limb loss. Artificial sensory feedback can address these deficits in real-world applications. Here, our findings provide an enhanced understanding of the optimal electrocutaneous stimulation parameters. They also demonstrate that low-cost, non-invasive electrocutaneous stimulation can serve as an appropriate alternative to the leading invasive methods. In all, electrocutaneous stimulation is the most accessible and intuitive method of artificial sensory feedback presented here. Its implementation will offer improved patient outcomes for those experiencing upper-limb loss and begin to restore the complete sensory experience of an intact hand. 33 ACKNOWLEDGMENTS This work was funded by NIH Award 1DP5OD029571-01 and the University of Utah Undergraduate Research Opportunities Program and Parent Fund – Undergraduate Research. 34 REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] K. Ziegler-Graham, E. J. MacKenzie, P. L. Ephraim, T. G. Travison, and R. Brookmeyer, “Estimating the Prevalence of Limb Loss in the United States: 2005 to 2050,” Arch. Phys. Med. Rehabil., vol. 89, no. 3, pp. 422–429, Mar. 2008, doi: 10.1016/j.apmr.2007.11.005. E. A. Biddiss and T. T. Chau, “Upper limb prosthesis use and abandonment: A survey of the last 25 years,” Prosthet. Orthot. Int., vol. 31, no. 3, pp. 236–257, Sep. 2007, doi: 10.1080/03093640600994581. E. Biddiss and T. Chau, “Upper-Limb Prosthetics: Critical Factors in Device Abandonment,” Am. J. Phys. Med. Rehabil., vol. 86, no. 12, pp. 977–987, Dec. 2007, doi: 10.1097/PHM.0b013e3181587f6c. D. Datta, K. Selvarajah, and N. Davey, “Functional outcome of patients with proximal upper limb deficiency–acquired and congenital,” Clin. Rehabil., vol. 18, no. 2, pp. 172–177, Mar. 2004, doi: 10.1191/0269215504cr716oa. M. Li et al., “Discrimination and Recognition of Phantom Finger Sensation Through Transcutaneous Electrical Nerve Stimulation,” Front. Neurosci., vol. 12, 2018, doi: 10.3389/fnins.2018.00283. L. Vargas, H. Shin, H. (Helen) Huang, Y. Zhu, and X. Hu, “Object stiffness recognition using haptic feedback delivered through transcutaneous proximal nerve stimulation,” J. Neural Eng., vol. 17, no. 1, p. 016002, Dec. 2019, doi: 10.1088/1741-2552/ab4d99. D. W. Tan, M. A. Schiefer, M. W. Keith, J. R. Anderson, J. Tyler, and D. J. Tyler, “A neural interface provides long-term stable natural touch perception,” Sci. Transl. Med., vol. 6, no. 257, p. 257ra138, Oct. 2014, doi: 10.1126/scitranslmed.3008669. S. Raspopovic et al., “Restoring Natural Sensory Feedback in Real-Time Bidirectional Hand Prostheses,” Sci. Transl. Med., vol. 6, no. 222, pp. 222ra19222ra19, Feb. 2014, doi: 10.1126/scitranslmed.3006820. D. M. Page et al., “Motor Control and Sensory Feedback Enhance Prosthesis Embodiment and Reduce Phantom Pain After Long-Term Hand Amputation,” Front. Hum. Neurosci., vol. 12, 2018, doi: 10.3389/fnhum.2018.00352. J. A. George et al., “Biomimetic sensory feedback through peripheral nerve stimulation improves dexterous use of a bionic hand,” Sci. Robot., vol. 4, no. 32, Jul. 2019, doi: 10.1126/scirobotics.aax2352. A. Ninu, S. Dosen, S. Muceli, F. Rattay, H. Dietl, and D. Farina, “Closed-Loop Control of Grasping With a Myoelectric Hand Prosthesis: Which Are the Relevant Feedback Variables for Force Control?,” IEEE Trans. Neural Syst. Rehabil. Eng., vol. 22, no. 5, pp. 1041–1052, Sep. 2014, doi: 10.1109/TNSRE.2014.2318431. E. I. Germany, E. J. Pino, and P. E. Aqueveque, “Myoelectric intuitive control and transcutaneous electrical stimulation of the forearm for vibrotactile sensation feedback applied to a 3D printed prosthetic hand,” in 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Aug. 2016, pp. 5046–5050. doi: 10.1109/EMBC.2016.7591861. 35 [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] M. Schiefer, D. Tan, S. M. Sidek, and D. J. Tyler, “Sensory feedback by peripheral nerve stimulation improves task performance in individuals with upper limb loss using a myoelectric prosthesis,” J. Neural Eng., vol. 13, no. 1, p. 016001, Dec. 2015, doi: 10.1088/1741-2560/13/1/016001. J. A. George, M. R. Brinton, P. C. Colgan, G. K. Colvin, S. J. Bensmaia, and G. A. Clark, “Intensity Discriminability of Electrocutaneous and Intraneural Stimulation Pulse Frequency in Intact Individuals and Amputees,” in 2020 42nd Annual International Conference of the IEEE Engineering in Medicine Biology Society (EMBC), Jul. 2020, pp. 3893–3896. doi: 10.1109/EMBC44109.2020.9176720. E. L. Graczyk, M. A. Schiefer, H. P. Saal, B. P. Delhaye, S. J. Bensmaia, and D. J. Tyler, “The neural basis of perceived intensity in natural and artificial touch,” Sci. Transl. Med., vol. 8, no. 362, pp. 362ra142-362ra142, Oct. 2016, doi: 10.1126/scitranslmed.aaf5187. C. Antfolk, M. D’Alonzo, B. Rosén, G. Lundborg, F. Sebelius, and C. Cipriani, “Sensory feedback in upper limb prosthetics,” Expert Rev. Med. Devices, vol. 10, no. 1, pp. 45–54, Jan. 2013, doi: 10.1586/erd.12.68. C. J. Kwolek and G. P. Clagett, “Changing demographics in patients with vascular disease,” J. Vasc. Surg., vol. 49, no. 2, pp. 528–531, Feb. 2009, doi: 10.1016/j.jvs.2008.11.075. R. C. M. van Kruijsdijk, Y. van der Graaf, R. H. H. Bemelmans, H. M. Nathoe, P. H. M. Peeters, and F. L. J. Visseren, “The relation between resting heart rate and cancer incidence, cancer mortality and all-cause mortality in patients with manifest vascular disease,” Cancer Epidemiol., vol. 38, no. 6, pp. 715–721, Dec. 2014, doi: 10.1016/j.canep.2014.09.004. J. W. P. Vernooij, Y. V. der Graaf, F. L. J. Visseren, and W. Spiering, “The Prevalence of Obesity-Related Hypertension and Risk for New Vascular Events in Patients With Vascular Diseases,” Obesity, vol. 20, no. 10, pp. 2118–2123, 2012, doi: 10.1038/oby.2012.76. W. J. Jeffcoate, “The Incidence of Amputation in Diabetes,” Acta Chir. Belg., vol. 105, no. 2, pp. 140–144, Jan. 2005, doi: 10.1080/00015458.2005.11679687. J. K. Gurney, J. Stanley, S. York, D. Rosenbaum, and D. Sarfati, “Risk of lower limb amputation in a national prevalent cohort of patients with diabetes,” Diabetologia, vol. 61, no. 3, pp. 626–635, Mar. 2018, doi: 10.1007/s00125-0174488-8. J. A. Barnes, M. A. Eid, M. A. Creager, and P. P. Goodney, “Epidemiology and Risk of Amputation in Patients With Diabetes Mellitus and Peripheral Artery Disease,” Arterioscler. Thromb. Vasc. Biol., vol. 40, no. 8, pp. 1808–1817, Aug. 2020, doi: 10.1161/ATVBAHA.120.314595. J. Poissy et al., “Pulmonary Embolism in Patients With COVID-19,” Circulation, vol. 142, no. 2, pp. 184–186, Jul. 2020, doi: 10.1161/CIRCULATIONAHA.120.047430. A. Scutelnic and M. R. Heldner, “Vascular Events, Vascular Disease and Vascular Risk Factors—Strongly Intertwined with COVID-19,” Curr. Treat. Options Neurol., vol. 22, no. 11, p. 40, 2020, doi: 10.1007/s11940-020-00648-y. 36 [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] H. K. Siddiqi, P. Libby, and P. M. Ridker, “COVID-19 – A vascular disease,” Trends Cardiovasc. Med., vol. 31, no. 1, pp. 1–5, Jan. 2021, doi: 10.1016/j.tcm.2020.10.005. T. R. Dillingham, L. E. Pezzin, and E. J. MacKenzie, “Limb amputation and limb deficiency: epidemiology and recent trends in the United States,” South. Med. J., vol. 95, no. 8, pp. 875–884, Aug. 2002. D. M. Desmond, “Coping, affective distress, and psychosocial adjustment among people with traumatic upper limb amputations,” J. Psychosom. Res., vol. 62, no. 1, pp. 15–21, Jan. 2007, doi: 10.1016/j.jpsychores.2006.07.027. J. Pillet and A. Didierjean-pillet, “AESTHETIC HAND PROSTHESIS: GADGET OR THERAPY? PRESENTATION OF A NEW CLASSIFICATION,” J. Hand Surg. Br. Eur. Vol., vol. 26, no. 6, pp. 523–528, Dec. 2001, doi: 10.1054/jhsb.2001.0658. D. M. Desmond and M. MacLachlan, “Prevalence and characteristics of phantom limb pain and residual limb pain in the long term after upper limb amputation,” Int. J. Rehabil. Res., vol. 33, no. 3, pp. 279–282, Sep. 2010, doi: 10.1097/MRR.0b013e328336388d. M. A. Hanley, D. M. Ehde, M. Jensen, J. Czerniecki, D. G. Smith, and L. R. Robinson, “Chronic Pain Associated with Upper-Limb Loss,” Am. J. Phys. Med. Rehabil. Assoc. Acad. Physiatr., vol. 88, no. 9, pp. 742–779, Sep. 2009, doi: 10.1097/PHM.0b013e3181b306ec. L. A. Wheaton, “Neurorehabilitation in upper limb amputation: understanding how neurophysiological changes can affect functional rehabilitation,” J. NeuroEngineering Rehabil., vol. 14, no. 1, p. 41, May 2017, doi: 10.1186/s12984017-0256-8. L. Resnik, M. Borgia, B. Silver, and J. Cancio, “Systematic Review of Measures of Impairment and Activity Limitation for Persons With Upper Limb Trauma and Amputation,” Arch. Phys. Med. Rehabil., vol. 98, no. 9, pp. 1863-1892.e14, Sep. 2017, doi: 10.1016/j.apmr.2017.01.015. K. Østlie, R. J. Franklin, O. H. Skjeldal, A. Skrondal, and P. Magnus, “Assessing Physical Function in Adult Acquired Major Upper-Limb Amputees by Combining the Disabilities of the Arm, Shoulder and Hand (DASH) Outcome Questionnaire and Clinical Examination,” Arch. Phys. Med. Rehabil., vol. 92, no. 10, pp. 1636– 1645, Oct. 2011, doi: 10.1016/j.apmr.2011.04.019. C. Gstoettner, S. Salminger, and O. C. Aszmann, “General Considerations on Upper Limb Amputation and Its Levels,” in Bionic Limb Reconstruction, O. C. Aszmann and D. Farina, Eds. Cham: Springer International Publishing, 2021, pp. 3–11. doi: 10.1007/978-3-030-60746-3_1. L. Resnik, M. Borgia, and M. Clark, “Function and Quality of Life of Unilateral Major Upper Limb Amputees: Effect of Prosthesis Use and Type,” Arch. Phys. Med. Rehabil., vol. 101, no. 8, pp. 1396–1406, Aug. 2020, doi: 10.1016/j.apmr.2020.04.003. K. Østlie, I. M. Lesjø, R. J. Franklin, B. Garfelt, O. H. Skjeldal, and P. Magnus, “Prosthesis use in adult acquired major upper-limb amputees: patterns of wear, prosthetic skills and the actual use of prostheses in activities of daily life,” 37 [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] Disabil. Rehabil. Assist. Technol., vol. 7, no. 6, pp. 479–493, Nov. 2012, doi: 10.3109/17483107.2011.653296. J. Uellendahl, “Body-Powered Prosthetic Systems,” in Bionic Limb Reconstruction, O. C. Aszmann and D. Farina, Eds. Cham: Springer International Publishing, 2021, pp. 27–35. doi: 10.1007/978-3-030-60746-3_4. F. Cordella et al., “Literature Review on Needs of Upper Limb Prosthesis Users,” Front. Neurosci., vol. 10, p. 209, 2016, doi: 10.3389/fnins.2016.00209. E. E. Desteli, Y. İmren, M. Erdoğan, G. Sarısoy, and S. Coşgun, “Comparison of upper limb amputees and lower limb amputees: a psychosocial perspective,” Eur. J. Trauma Emerg. Surg., vol. 40, no. 6, pp. 735–739, Dec. 2014, doi: 10.1007/s00068-014-0418-3. M. Tsakiris and P. Haggard, “The Rubber Hand Illusion Revisited: Visuotactile Integration and Self-Attribution,” J. Exp. Psychol. Hum. Percept. Perform., vol. 31, no. 1, pp. 80–91, 2005, doi: 10.1037/0096-1523.31.1.80. E. C. Lape, P. Hudak, A. M. Davis, and J. N. Katz, “Body-Self Unity With a New Hip or Knee: Understanding Total Joint Replacement Within an Embodiment Framework,” ACR Open Rheumatol., vol. 1, no. 2, pp. 90–96, 2019, doi: 10.1002/acr2.1014. A. Cierpka, K. Malek, and M. Horodeńska, “Original article What does it mean to live after heart transplantation? The lived experience of heart transplant recipients. A qualitative study,” Health Psychol. Rep., vol. 3, Mar. 2015, doi: 10.5114/hpr.2015.49678. P. U. Dijkstra, J. H. B. Geertzen, R. Stewart, and C. P. van der Schans, “Phantom pain and risk factors: a multivariate analysis,” J. Pain Symptom Manage., vol. 24, no. 6, pp. 578–585, Dec. 2002, doi: 10.1016/s0885-3924(02)00538-9. C. M. Fraser, P. W. Halligan, I. H. Robertson, and S. G. Kirker, “Characterising phantom limb phenomena in upper limb amputees,” Prosthet. Orthot. Int., vol. 25, no. 3, pp. 235–242, Dec. 2001, doi: 10.1080/03093640108726607. H. Flor, “Phantom-limb pain: characteristics, causes, and treatment,” Lancet Neurol., vol. 1, no. 3, pp. 182–189, Jul. 2002, doi: 10.1016/S14744422(02)00074-1. P. S. Mckechnie and A. John, “Anxiety and depression following traumatic limb amputation: A systematic review,” Injury, vol. 45, no. 12, pp. 1859–1866, Dec. 2014, doi: 10.1016/j.injury.2014.09.015. N. A. Hashim, N. A. Abd Razak, N. A. Abu Osman, and H. Gholizadeh, “Improvement on upper limb body-powered prostheses (1921–2016): A systematic review,” Proc. Inst. Mech. Eng. [H], vol. 232, no. 1, pp. 3–11, Jan. 2018, doi: 10.1177/0954411917744585. A. Calado, F. Soares, and D. Matos, “A Review on Commercially Available Anthropomorphic Myoelectric Prosthetic Hands, Pattern-Recognition-Based Microcontrollers and sEMG Sensors used for Prosthetic Control,” in 2019 IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC), Apr. 2019, pp. 1–6. doi: 10.1109/ICARSC.2019.8733629. A. D. Roche, H. Rehbaum, D. Farina, and O. C. Aszmann, “Prosthetic Myoelectric Control Strategies: A Clinical Perspective,” Curr. Surg. Rep., vol. 2, 38 no. 3, p. 44, Jan. 2014, doi: 10.1007/s40137-013-0044-8. [50] S. Salminger et al., “Functional Outcome Scores With Standard Myoelectric Prostheses in Below-Elbow Amputees,” Am. J. Phys. Med. Rehabil., vol. 98, no. 2, pp. 125–129, Feb. 2019, doi: 10.1097/PHM.0000000000001031. [51] S. Salminger et al., “Current rates of prosthetic usage in upper-limb amputees have innovations had an impact on device acceptance?,” Disabil. Rehabil., pp. 1– 12, Dec. 2020, doi: 10.1080/09638288.2020.1866684. [52] N. Parajuli et al., “Real-Time EMG Based Pattern Recognition Control for Hand Prostheses: A Review on Existing Methods, Challenges and Future Implementation,” Sensors, vol. 19, no. 20, Art. no. 20, Jan. 2019, doi: 10.3390/s19204596. [53] “LUKE Arm Detail Page – Mobius Bionics.” https://www.mobiusbionics.com/luke-arm/ (accessed Sep. 17, 2021). [54] S. M. Engdahl, C. A. Chestek, B. Kelly, A. Davis, and D. H. Gates, “Factors associated with interest in novel interfaces for upper limb prosthesis control,” PLoS ONE, vol. 12, no. 8, p. e0182482, Aug. 2017, doi: 10.1371/journal.pone.0182482. [55] S. M. Engdahl, B. P. Christie, B. Kelly, A. Davis, C. A. Chestek, and D. H. Gates, “Surveying the interest of individuals with upper limb loss in novel prosthetic control techniques,” J. NeuroEngineering Rehabil., vol. 12, no. 1, 2015, doi: 10.1186/s12984-015-0044-2. [56] S. Lewis, M. F. Russold, H. Dietl, and E. Kaniusas, “User demands for sensory feedback in upper extremity prostheses,” in 2012 IEEE International Symposium on Medical Measurements and Applications Proceedings, May 2012, pp. 1–4. doi: 10.1109/MeMeA.2012.6226669. [57] M. Markovic et al., “The clinical relevance of advanced artificial feedback in the control of a multi-functional myoelectric prosthesis,” J. NeuroEngineering Rehabil., vol. 15, no. 1, p. 28, Mar. 2018, doi: 10.1186/s12984-018-0371-1. [58] F. Clemente et al., “Intraneural sensory feedback restores grip force control and motor coordination while using a prosthetic hand,” J. Neural Eng., vol. 16, no. 2, p. 026034, Feb. 2019, doi: 10.1088/1741-2552/ab059b. [59] E. D’Anna et al., “A closed-loop hand prosthesis with simultaneous intraneural tactile and position feedback,” Sci. Robot., Feb. 2019, doi: 10.1126/scirobotics.aau8892. [60] D. M. Page et al., “Discriminability of multiple cutaneous and proprioceptive hand percepts evoked by intraneural stimulation with Utah slanted electrode arrays in human amputees,” J. NeuroEngineering Rehabil., vol. 18, no. 1, p. 12, Jan. 2021, doi: 10.1186/s12984-021-00808-4. [61] G. F. Shannon, “A myoelectrically-controlled prosthesis with sensory feedback,” Med. Biol. Eng. Comput., vol. 17, no. 1, pp. 73–80, Jan. 1979, doi: 10.1007/BF02440956. [62] S. Wendelken et al., “Restoration of motor control and proprioceptive and cutaneous sensation in humans with prior upper-limb amputation via multiple Utah Slanted Electrode Arrays (USEAs) implanted in residual peripheral arm nerves,” J. NeuroEngineering Rehabil., vol. 14, no. 1, p. 121, Nov. 2017, doi: 39 [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] 10.1186/s12984-017-0320-4. A. R. Sobinov and S. J. Bensmaia, “The neural mechanisms of manual dexterity,” Nat. Rev. Neurosci., vol. 22, no. 12, Art. no. 12, Dec. 2021, doi: 10.1038/s41583021-00528-7. G. Rognini et al., “Multisensory bionic limb to achieve prosthesis embodiment and reduce distorted phantom limb perceptions,” J. Neurol. Neurosurg. Psychiatry, vol. 90, no. 7, pp. 833–836, Jul. 2019, doi: 10.1136/jnnp-2018318570. S. Gholinezhad, S. Dosen, and D. Jakob, “Electrotactile feedback outweighs natural feedback in sensory integration during control of grasp force,” J. Neural Eng., vol. 18, no. 5, p. 056024, Oct. 2021, doi: 10.1088/1741-2552/ac1fce. C. Dietrich et al., “Leg Prosthesis With Somatosensory Feedback Reduces Phantom Limb Pain and Increases Functionality,” Front. Neurol., vol. 9, p. 270, 2018, doi: 10.3389/fneur.2018.00270. C. Dietrich et al., “Sensory feedback prosthesis reduces phantom limb pain: Proof of a principle,” Neurosci. Lett., vol. 507, no. 2, pp. 97–100, Jan. 2012, doi: 10.1016/j.neulet.2011.10.068. P. D. Marasco, K. Kim, J. E. Colgate, M. A. Peshkin, and T. A. Kuiken, “Robotic touch shifts perception of embodiment to a prosthesis in targeted reinnervation amputees,” Brain, vol. 134, no. 3, pp. 747–758, Mar. 2011, doi: 10.1093/brain/awq361. M. Gadhvi and M. Waseem, “Physiology, Sensory System,” in StatPearls, Treasure Island (FL): StatPearls Publishing, 2021. Accessed: Sep. 29, 2021. [Online]. Available: http://www.ncbi.nlm.nih.gov/books/NBK547656/ H. P. Saal and S. J. Bensmaia, “Biomimetic approaches to bionic touch through a peripheral nerve interface,” Neuropsychologia, vol. 79, pp. 344–353, Dec. 2015, doi: 10.1016/j.neuropsychologia.2015.06.010. M. Paré, R. Elde, J. E. Mazurkiewicz, A. M. Smith, and F. L. Rice, “The Meissner Corpuscle Revised: A Multiafferented Mechanoreceptor with Nociceptor Immunochemical Properties,” J. Neurosci., vol. 21, no. 18, pp. 7236–7246, Sep. 2001, doi: 10.1523/JNEUROSCI.21-18-07236.2001. Y. Roudaut, A. Lonigro, B. Coste, J. Hao, P. Delmas, and M. Crest, “Touch sense,” Channels, vol. 6, no. 4, pp. 234–245, Jul. 2012, doi: 10.4161/chan.22213. P. Herberts, “Ideas on sensory feedback in hand prostheses | O&P Virtual Library,” Prosthet. Orthot. Int., vol. 3, no. 3, pp. 157–162, 1979. R. G. E. Clement, K. E. Bugler, and C. W. Oliver, “Bionic prosthetic hands: A review of present technology and future aspirations,” The Surgeon, vol. 9, no. 6, pp. 336–340, Dec. 2011, doi: 10.1016/j.surge.2011.06.001. N. Jiang, S. Dosen, K.-R. Muller, and D. Farina, “Myoelectric Control of Artificial Limbs—Is There a Need to Change Focus? [In the Spotlight],” IEEE Signal Process. Mag., vol. 29, no. 5, pp. 152–150, Sep. 2012, doi: 10.1109/MSP.2012.2203480. D. J. DiLorenzo, “Development of a chronically implanted microelectrode array for intraneural electrical stimulation for prosthetic sensory feedback,” Thesis, Massachusetts Institute of Technology, 1999. Accessed: Sep. 16, 2021. [Online]. 40 [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] Available: https://dspace.mit.edu/handle/1721.1/9329 M. B. Christensen, S. M. Pearce, N. M. Ledbetter, D. J. Warren, G. A. Clark, and P. A. Tresco, “The foreign body response to the Utah Slant Electrode Array in the cat sciatic nerve,” Acta Biomater., vol. 10, no. 11, pp. 4650–4660, Nov. 2014, doi: 10.1016/j.actbio.2014.07.010. A. Sharma et al., “Long termin vitrofunctional stability and recording longevity of fully integrated wireless neural interfaces based on the Utah Slant Electrode Array,” J. Neural Eng., vol. 8, no. 4, p. 045004, Jul. 2011, doi: 10.1088/17412560/8/4/045004. R. Ackerley, H. Backlund Wasling, M. Ortiz-Catalan, R. Brånemark, and J. Wessberg, “Case Studies in Neuroscience: Sensations elicited and discrimination ability from nerve cuff stimulation in an amputee over time,” J. Neurophysiol., vol. 120, no. 1, pp. 291–295, Jul. 2018, doi: 10.1152/jn.00909.2017. D. K. Leventhal and D. M. Durand, “Subfascicle Stimulation Selectivity with the Flat Interface Nerve Electrode,” Ann. Biomed. Eng., vol. 31, no. 6, pp. 643–652, Jun. 2003, doi: 10.1114/1.1569266. E. L. Graczyk, B. P. Christie, Q. He, D. J. Tyler, and S. J. Bensmaia, “Frequency shapes the quality of tactile percepts evoked through electrical stimulation of the nerves,” bioRxiv, p. 2020.08.24.263822, Aug. 2020, doi: 10.1101/2020.08.24.263822. F. A. Geldard, “The Perception of Mechanical Vibration: I. History of a Controversy,” J. Gen. Psychol., vol. 22, no. 2, pp. 243–269, Apr. 1940, doi: 10.1080/00221309.1940.9710033. C. E. Sherrick, “The art of tactile communication,” Am. Psychol., vol. 30, no. 3, pp. 353–360, Mar. 1975, doi: 10.1037/0003-066X.30.3.353. V. O. Knudsen, “‘Hearing’ with the Sense of Touch,” J. Gen. Psychol., vol. 1, no. 2, pp. 320–352, Apr. 1928, doi: 10.1080/00221309.1928.9920128. L. Burns, “Music in Deaf Education,” Music J., vol. 23, no. 8, p. 42, 1965. E. Foulke, “The Locus Dimension as a Basis for Electrocutaneous Communication,” J. Psychol., 1963, doi: 10.1080/00223980.1964.9916695 E. Foulke and A. A. Brodbeck, “Transmission of Morse Code by Electrocutaneous Stimulation,” Psychol. Rec., vol. 18, no. 4, pp. 617–622, Oct. 1968, doi: 10.1007/BF03393813. G. B. Rollman, “Electrocutaneous stimulation: Psychometric functions and temporal integration,” Percept. Psychophys., vol. 5, no. 5, pp. 289–293, Sep. 1969, doi: 10.3758/BF03209565. T. Tashiro and A. Higashiyama, “The perceptual properties of electrocutaneous stimulation: Sensory quality, subjective intensity, and intensity-duration relation,” Percept. Psychophys., vol. 30, no. 6, pp. 579–586, Nov. 1981, doi: 10.3758/BF03202013. R. Gibson, “Electrical Stimulation of Pain and Touch Systems,” Nature, vol. 199, pp. 307–308, Jul. 1963. R. E. Prior, “Electrocutaneous Feedback for Artificial Limbs Summary Progress Report, February 1, 1974 through July 31, 1975,” p. 35. R. N. Scott, R. H. Brittain, R. R. Caldwell, A. B. Cameron, and V. A. Dunfield, 41 “Sensory-feedback system compatible with myoelectric control,” Med. Biol. Eng. Comput., vol. 18, no. 1, pp. 65–69, Jan. 1980, doi: 10.1007/BF02442481. [93] C. Almström, A. Anani, P. Herberts, and L. Körner, “Electrical stimulation and myoelectric control. A theoretical and applied study relevant to prosthesis sensory feedback,” Med. Biol. Eng. Comput., vol. 19, no. 5, pp. 645–653, Sep. 1981, doi: 10.1007/BF02442780. [94] A. Y. J. Szeto and F. A. Saunders, “Electrocutaneous Stimulation for Sensory Communication in Rehabilitation Engineering,” IEEE Trans. Biomed. Eng., vol. BME-29, no. 4, pp. 300–308, Apr. 1982, doi: 10.1109/TBME.1982.324948. [95] D. S. Childress, “Closed-loop control in prosthetic systems: Historical perspective,” Ann. Biomed. Eng., vol. 8, no. 4–6, pp. 293–303, Jul. 1980, doi: 10.1007/BF02363433. [96] J. L. Segil, I. Cuberovic, E. L. Graczyk, R. F. ff Weir, and D. Tyler, “Combination of Simultaneous Artificial Sensory Percepts to Identify Prosthetic Hand Postures: A Case Study,” Sci. Rep., vol. 10, no. 1, p. 6576, Apr. 2020, doi: 10.1038/s41598-020-62970-4. [97] P. Svensson, U. Wijk, A. Björkman, and C. Antfolk, “A review of invasive and non-invasive sensory feedback in upper limb prostheses,” Expert Rev. Med. Devices, vol. 14, no. 6, pp. 439–447, Jun. 2017, doi: 10.1080/17434440.2017.1332989. [98] B. Geng, K. Yoshida, L. Petrini, and W. Jensen, “Evaluation of sensation evoked by electrocutaneous stimulation on forearm in nondisabled subjects,” J. Rehabil. Res. Dev., vol. 49, no. 2, p. 297, 2012, doi: 10.1682/JRRD.2010.09.0187. [99] L. Seminara et al., “Electronic skin and electrocutaneous stimulation to restore the sense of touch in hand prosthetics,” in 2017 IEEE International Symposium on Circuits and Systems (ISCAS), May 2017, pp. 1–4. doi: 10.1109/ISCAS.2017.8050264. [100] Y. Abbass, M. Saleh, S. Dosen, and M. Valle, “Embedded Electrotactile Feedback System for Hand Prostheses using Matrix Electrode and Electronic Skin,” IEEE Trans. Biomed. Circuits Syst., pp. 1–1, 2021, doi: 10.1109/TBCAS.2021.3107723. [101] J. Dideriksen, M. Markovic, S. Lemling, D. Farina, and S. Dosen, “Electrotactile and Vibrotactile Feedback Enable Similar Performance in Psychometric Tests and Closed-loop Control,” IEEE Trans. Haptics, pp. 1–1, 2021, doi: 10.1109/TOH.2021.3117628. [102] M. Isakovic, J. Malesevic, M. Kostic, S. Dosen, and M. Strbac, “The impact of size and position of reference electrode on the localization of biphasic electrotactile stimulation on the fingertips,” IEEE Trans. Haptics, pp. 1–1, 2022, doi: 10.1109/TOH.2022.3141187. [103] J. Zhang et al., “Modulation Ranges of Different Sensations for Coding Electrically Evoked Tactile Sensory Feedback,” May 2021, doi: 10.36227/techrxiv.14693934.v1. [104] G. K. Patel, S. Dosen, C. Castellini, and D. Farina, “Multichannel electrotactile feedback for simultaneous and proportional myoelectric control,” J. Neural Eng., vol. 13, no. 5, p. 056015, Sep. 2016, doi: 10.1088/1741-2560/13/5/056015. 42 [105] D. G. Buma, J. R. Buitenweg, and P. H. Veltink, “Intermittent Stimulation Delays Adaptation to Electrocutaneous Sensory Feedback,” IEEE Trans. Neural Syst. Rehabil. Eng., vol. 15, no. 3, pp. 435–441, Sep. 2007, doi: 10.1109/TNSRE.2007.903942. [106] E. L. Graczyk, B. P. Delhaye, M. A. Schiefer, S. J. Bensmaia, and D. J. Tyler, “Sensory adaptation to electrical stimulation of the somatosensory nerves,” J. Neural Eng., vol. 15, no. 4, p. 046002, May 2018, doi: 10.1088/17412552/aab790. [107] J. Caldwell, “A High-Voltage Bidirectional Current Source,” p. 9. [108] D. R. Merrill, M. Bikson, and J. G. R. Jefferys, “Electrical stimulation of excitable tissue: design of efficacious and safe protocols,” J. Neurosci. Methods, vol. 141, no. 2, pp. 171–198, Feb. 2005, doi: 10.1016/j.jneumeth.2004.10.020. [109] M. A. Muniak, S. Ray, S. S. Hsiao, J. F. Dammann, and S. J. Bensmaia, “The neural coding of stimulus intensity: Linking the population response of mechanoreceptive afferents with psychophysical behavior (Journal of Neuroscience (2007) (11687-11699)),” J. Neurosci., vol. 28, no. 7, Feb. 2008, Accessed: Apr. 27, 2021. [Online]. Available: https://jhu.pure.elsevier.com/en/publications/the-neural-coding-of-stimulusintensity-linking-the-population-re-10 [110] S. Musall, W. von der Behrens, J. M. Mayrhofer, B. Weber, F. Helmchen, and F. Haiss, “Tactile frequency discrimination is enhanced by circumventing neocortical adaptation,” Nat. Neurosci., vol. 17, no. 11, pp. 1567–1573, Nov. 2014, doi: 10.1038/nn.3821. [111] A. Y. J. Szeto, J. Lyman, and R. E. Prior, “Electrocutaneous Pulse Rate and Pulse Width Psychometric Functions for Sensory Communications,” Hum. Factors, vol. 21, no. 2, pp. 241–249, Apr. 1979, doi: 10.1177/001872087902100212. [112] Gös. Ekman, “Weber’s Law and Related Functions,” J. Psychol., vol. 47, no. 2, pp. 343–352, Apr. 1959, doi: 10.1080/00223980.1959.9916336. [113] L. Vargas, H. Huang, Y. Zhu, and X. Hu, “Object Recognition via Evoked Sensory Feedback during Control of a Prosthetic Hand,” IEEE Robot. Autom. Lett., vol. 7, no. 1, pp. 207–214, Jan. 2022, doi: 10.1109/LRA.2021.3122897. [114] F. Khabbaz, A. A. Goldenberg, and J. Drake, “Force Discrimination Ability of Human Hand near Absolute Threshold for the Design of Force Feedback Systems in Teleoperations,” Presence Teleoperators Virtual Environ., vol. 25, Jun. 2016, doi: 10.1162/PRES_a_00245. [115] R. Höver, M. D. Luca, and M. Harders, “User-based evaluation of data-driven haptic rendering,” ACM Trans. Appl. Percept., vol. 8, no. 1, pp. 1–23, Oct. 2010, doi: 10.1145/1857893.1857900. [116] F. Ranieri et al., “Sensorimotor integration within the primary motor cortex by selective nerve fascicle stimulation,” J. Physiol., vol. n/a, no. n/a, doi: 10.1113/JP282259. [117] C. L. Hughes and R. A. Gaunt, “Changes in interpulse spacing changes tactile perception of microstimulation in human somatosensory cortex,” in 2021 10th International IEEE/EMBS Conference on Neural Engineering (NER), May 2021, pp. 660–663. doi: 10.1109/NER49283.2021.9441292. 43 [118] G. Valle et al., “Biomimetic Intraneural Sensory Feedback Enhances Sensation Naturalness, Tactile Sensitivity, and Manual Dexterity in a Bidirectional Prosthesis,” Neuron, vol. 100, no. 1, pp. 37-45.e7, Oct. 2018, doi: 10.1016/j.neuron.2018.08.033. [119] M. Ortiz-Catalan, J. Wessberg, E. Mastinu, A. Naber, and R. Brånemark, “Patterned Stimulation of Peripheral Nerves Produces Natural Sensations With Regards to Location but Not Quality,” IEEE Trans. Med. Robot. Bionics, vol. 1, no. 3, pp. 199–203, Aug. 2019, doi: 10.1109/TMRB.2019.2931758. [120] M. W. Brault, J. Hootman, C. G. Helmick, K. A. Theis, and B. S. Armour, “Prevalence and most common causes of disability among adults--United States, 2005,” Morb. Mortal. Wkly. Rep., vol. 58, no. 16, pp. 421–426, May 2009. [121] S. Koton et al., “Stroke Incidence and Mortality Trends in US Communities, 1987 to 2011,” JAMA, vol. 312, no. 3, pp. 259–268, Jul. 2014, doi: 10.1001/jama.2014.7692. [122] D. K. Sommerfeld, E. U.-B. Eek, A.-K. Svensson, L. W. Holmqvist, and M. H. von Arbin, “Spasticity After Stroke,” Stroke, vol. 35, no. 1, pp. 134–139, Jan. 2004, doi: 10.1161/01.STR.0000105386.05173.5E. [123] L. M. Carey and T. A. Matyas, “Frequency of discriminative sensory loss in the hand after stroke in a rehabilitation setting,” J. Rehabil. Med., vol. 43, no. 3, pp. 257–263, Feb. 2011, doi: 10.2340/16501977-0662. [124] S. F. Tyson, M. Hanley, J. Chillala, A. B. Selley, and R. C. Tallis, “Sensory Loss in Hospital-Admitted People With Stroke: Characteristics, Associated Factors, and Relationship With Function,” Neurorehabil. Neural Repair, vol. 22, no. 2, pp. 166–172, Mar. 2008, doi: 10.1177/1545968307305523. [125] C. E. Winward, P. W. Halligan, and D. T. Wade, “The Rivermead Assessment of Somatosensory Performance (RASP): standardization and reliability data,” Clin. Rehabil., vol. 16, no. 5, pp. 523–533, Aug. 2002, doi: 10.1191/0269215502cr522oa. [126] S. Hillier, M. Immink, and D. Thewlis, “Assessing Proprioception: A Systematic Review of Possibilities,” Neurorehabil. Neural Repair, vol. 29, no. 10, pp. 933– 949, Nov. 2015, doi: 10.1177/1545968315573055. [127] J. Hallam, “Haptic mirror therapy glove: aiding the treatment of a paretic limb after a stroke,” in Adjunct Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2015 ACM International Symposium on Wearable Computers, New York, NY, USA, Sep. 2015, pp. 459–464. doi: 10.1145/2800835.2801648. [128] J. C. Kattenstroth, T. Kalisch, M. Sczesny-Kaiser, W. Greulich, M. Tegenthoff, and H. R. Dinse, “Daily repetitive sensory stimulation of the paretic hand for the treatment of sensorimotor deficits in patients with subacute stroke: RESET, a randomized, sham-controlled trial,” BMC Neurol., vol. 18, no. 1, p. 2, Jan. 2018, doi: 10.1186/s12883-017-1006-z. [129] “Reality Labs,” Facebook Technology. https://tech.fb.com/ar-vr/ (accessed Feb. 01, 2022). [130] M. Salvato et al., “Data-driven sparse skin stimulation can convey social touch information to humans,” IEEE Trans. Haptics, pp. 1–1, 2021, doi: 44 [131] [132] [133] [134] 10.1109/TOH.2021.3129067. X. L. Cang and A. Israr, “Communicating Socio-Emotional Sentiment Through Haptic Messages,” p. 3. E. D. Z. Chase, A. Israr, P. Preechayasomboon, S. Sykes, A. Gupta, and J. Hartcher-O’Brien, “Learning Vibes: Communication Bandwidth of a Single Wrist-Worn Vibrotactile Actuator,” in 2021 IEEE World Haptics Conference (WHC), Montreal, QC, Canada, Jul. 2021, pp. 421–426. doi: 10.1109/WHC49131.2021.9517208. E. Pezent et al., “Explorations of Wrist Haptic Feedback for AR/VR Interactions with Tasbi,” in Extended Abstracts of the 2020 CHI Conference on Human Factors in Computing Systems, Honolulu HI USA, Apr. 2020, pp. 1–4. doi: 10.1145/3334480.3383151. A. Gupta, M. Samad, K. Kin, P. O. Kristensson, and H. Benko, “Investigating Remote Tactile Feedback for Mid-Air Text-Entry in Virtual Reality,” in 2020 IEEE International Symposium on Mixed and Augmented Reality (ISMAR), Porto de Galinhas, Brazil, Nov. 2020, pp. 350–360. doi: 10.1109/ISMAR50242.2020.00062. 45 Name of Candidate: Abigail R. Citterman Date of Submission: March 20, 2022 46