Creating a neuroprosthesis for active tactile exploration of textures

Abstract

39Intracortical microstimulation (ICMS) of the primary somatosensory cortex (S1) can 40 produce percepts that mimic somatic sensation and thus has potential as an approach to 41 sensorize prosthetic limbs. However, it is not known whether ICMS could recreate active texture 42 exploration-the ability to infer information about object texture by using one's fingertips to scan 43 a surface. Here we show that ICMS of S1 can convey information about the spatial frequencies 44 of invisible virtual gratings through a process of active tactile exploration. Two rhesus monkeys 45 scanned pairs of visually identical screen objects with the fingertip of a hand avatar, controlled 46 via a joystick and later via a brain-machine interface, to find the one with denser virtual gratings. 47The gratings consisted of evenly spaced ridges that were signaled through ICMS pulses 48 generated when the avatar's fingertip crossed each ridge. The monkeys learned to interpret 49 these ICMS patterns evoked by the interplay of their voluntary movements and the virtual 50 textures of each object. Discrimination accuracy across a range of grating densities followed 51Weber's law of just-noticeable differences (JND), a finding that matches normal cutaneous 52 sensation. Moreover, one monkey developed an active scanning strategy where avatar velocity 53 was integrated with the ICMS pulses to interpret the texture information. We propose that this 54 approach could equip upper-limb neuroprostheses with direct access to texture features 55 acquired during active exploration of natural objects. 57Sensory neuroprostheses offer the promise of restoring perceptual function to people 59 with impaired sensation 1,2 . In such devices, diminished sensory modalities (e.g., hearing 3 , 60 vision 4,5 , or cutaneous touch 6-8 ) are reenacted through streams of artificial input to the nervous 61 system, typically using electrical stimulation of nerve fibers in the periphery or neurons in the 62 central nervous system. Restored cutaneous touch, in particular, would be of great benefit for 63 the users of upper-limb prostheses, who place a high priority on the ability to perform functions 64 without the necessity to constantly engage visual attention 9 . This could be achieved through the 65 addition of artificial somatosensory channels to the prosthetic device 1 . Such an approach would 66 endow persons suffering from limb loss 10-12 , paralysis 1,13 or somatosensory deficits with the 67 ability to perform active tactile exploration of their physical environment and aid in dexterous 68 object manipulation [14][15][16][17] . 69Previously we demonstrated that motor and sensory functions could be simultaneously 70 enacted though a bidirectional neuroprosthetic system, called a brain-machine-brain interface 71 (BMBI) 18 . In that demonstration, the active exploration enabled by our BMBI-driven 72 neuroprosthesis used a limited and fixed set of ICMS temporal patterns to generate artificial 73 sensory inputs that mimicked the sense of flutter-vibration. However, it r...

Sputtered ruthenium oxide coatings for neural stimulation and recording electrodes

“…Recording and stimulation of neural activity are employed in devices intended to restore lost or impaired neurological function in humans. Applications may include brain machine interfaces (BMIs) for volitional control of extracorporeal prostheses, intracortical vision prostheses, auditory brain stem prostheses, and deep‐brain stimulation 1‐6 . These devices require electrodes that are characterized by low impedance for recording and a high reversible charge injection for stimulation 7 .…”

Brain-machine interface learning is facilitated by specific patterning of distributed cortical feedback

“…This choice ensured that the optogenetic feedback delivered to S1 was the sole source of sensory information about the cursor position available to the animal during the task. This is in contrast to most previous closed-loop BMI studies, in which ongoing visual feedback of the neuroprosthesis was always available for adjusting motor control (Flesher et al, 2019;O'Doherty et al, 2019). Yet in a recent study, mice were trained to condition motor cortex neurons in an operant way while receiving ratemodulated pulses of optogenetic stimulation in S1, in the absence of visual information (Prsa et al, 2017).…”

“…Current research aimed at integrating somatosensory feedback in a cortical brain-machine interface relies on invasive techniques of recording and stimulation in awake behaving animals. Pioneering teams are developing prototypes in non-human primates as well as human participants (Flesher et al, 2019;O'Doherty et al, 2019). Here, we have developed a novel brain-machine interface tailored to the mouse whisker system, a sensorimotor loop that has been described in a very comprehensive way, from the cellular to the network scale (Diamond and Arabzadeh, 2013;Petersen, 2019).…”

“…In the context of brain-machine interfaces, proprioceptive and touch-like feedback originating from the prosthesis could improve its control. Indeed, recent brain-machine interfaces that have incorporated an afferent somatosensory component have achieved faster prosthesis control (Flesher et al, 2019) and elicited texture-like percepts that cannot be obtained through visual feedback alone (O'Doherty et al, 2019). Thus, for building efficient brain-machine interfaces, a widespread hypothesis is that the feedback should convey somatosensory information about ongoing consequences of the motor control (Bensmaia and Miller, 2014).…”

Injecting Information into the Mammalian Cortex: Progress, Challenges, and Promise

“…ICMS in S1 thus can be biomimetic for vibration frequency, pressure magnitude, and skin location. Beyond these passively perceived features, monkeys also have discriminated grating textures mimicked by delivering an ICMS pulse in S1 each time the fingertip of an actively exploring avatar hand (controlled by the monkey with either a joystick or a BCI) encountered a virtual ridge at spatial frequencies of 0.5 to 4.0/cm, which requires integrating the frequency of ICMS pulses with knowledge of the speed of hand motion (O’Doherty and others 2019). In this texture discrimination task the monkeys’ performance was above chance from the outset, but improved further with experience.…”