Introduction

Blindness affects millions of individuals worldwide and presents numerous challenges in daily life[1](https://www.nature.com/articles/s41598-025-21720-0#ref-CR1 “World Health Organization Blindness and Vision Impairment. Retrieved August 13, 2024, from https://www.who.int/news-room/fact-sheets/detail/blindness-and-visual-impairment

(2023).“). For people with limited vision, activities fundamental to everyday tasks, such as eating, dressing, and manipulating objects, become more complex due to the absence of visual guidance and adjustment of movements. This constraint inherently limits their capacity to fully participate in activities that sighted individuals often take for granted[2](https://www.nature.com/articles/s41598-025-21720-0#ref-CR2 “Langelaan, M. et al. Impact of visual impairment on quality of life: A comparison with quality of life in the general population and with other chronic conditions. OphthalmicEpidemiol. 14, 119–126. https://doi.org/10.1080/09286580601139212

(2007).“). Recent advancements in assistive Quality-of-life technology, such as obstacle detection devices for the body and the white cane, as well as audio-based navigation apps, have empowered many people with limited vision to achieve greater independence[2](https://www.nature.com/articles/s41598-025-21720-0#ref-CR2 “Langelaan, M. et al. Impact of visual impairment on quality of life: A comparison with quality of life in the general population and with other chronic conditions. OphthalmicEpidemiol. 14, 119–126. https://doi.org/10.1080/09286580601139212

(2007).“),[3](https://www.nature.com/articles/s41598-025-21720-0#ref-CR3 “Muhsin, Z. J., Qahwaji, R. & Ghanchi, F. Review of substitutive assistive tools and technologies for people with visual impairments: recent advancements and prospects. J. Multimodal. User Interfaces 18, 135–156. https://doi.org/10.1007/s12193-023-00427-4

(2023).“). Ultimately, by prioritizing accessibility and innovative solutions, society can significantly enhance the autonomy and quality of life for those with visual conditions.

Especially navigating everyday tasks, like grocery shopping, identifying food, and grasping desired objects, can be particularly challenging for people with limited vision due to their reliance on non-visual senses. The grasping movements representing a central difficulty as vision is the primary sensory system for spatial tasks like grasping4. People with limited vision must rely on alternative senses such as touch, hearing, proprioception, and spatial memory to compensate the lack of visual input[5](#ref-CR5 “Thinus-Blanc, C. & Gaunet, F. Representation of space in blind persons: Vision as a spatial sense?. Psychol. Bull. 121(1), 20–42. https://doi.org/10.1037/0033-2909.121.1.20

(1997).“),[6](#ref-CR6 “Pardhan, S., Gonzalez-Alvarez, C. & Subramanian, A. How does the presence and duration of central visual impairment affect reaching and grasping movements?. Ophthalmic Physiol. Opt. 31, 233–239. https://doi.org/10.1111/j.1475-1313.2010.00819.x

(2011).“),[7](#ref-CR7 “Yatani, K., Banovic, N. & Truong, K. N. SpaceSense: Representing Geographical Information to Visually Impaired People Using Spatial Tactile Feedback. In: CHI ‘12: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. 415–424. https://doi.org/10.1145/2207676.2207734

(2012).“),[8](#ref-CR8 “Kaspar, K., König, S., Schwandt, J. & König, P. The experience of new sensorimotor contingencies by sensory augmentation. Conscious. Cogn. 28, 47–63. https://doi.org/10.1016/j.concog.2014.06.006

(2014).“),[9](https://www.nature.com/articles/s41598-025-21720-0#ref-CR9 “Kärcher, S. M., Fenzlaff, S., Hartmann, D., Nagel, S. K. & König, P. Sensory augmentation for the blind. Front. Human Neurosci. https://doi.org/10.3389/fnhum.2012.00037

(2012).“). So, everyday actions are more complex, which highlights the need for accessible design and assistive tools.

Consequently, individuals with visual conditions, despite possessing the physical capacity for grasping, they face significant challenges due to lacking this essential sensory information4,[10](https://www.nature.com/articles/s41598-025-21720-0#ref-CR10 “Stone, K. D. & Gonzalez, C. L. The contributions of vision and haptics to reaching and grasping. Front. Psychol. https://doi.org/10.3389/fpsyg.2015.01403

(2015).“). The developmental foundations of tactile and multimodal-guided reaching in visually conditioned individuals are complex and not yet fully understood. Work by Gori et al.[11](https://www.nature.com/articles/s41598-025-21720-0#ref-CR11 “Gori, M., Campus, C., Signorini, S., Rivara, E. & Bremner, A. J. Multisensory spatial perception in visually impaired infants. Curr. Biol. CB 31(22), 5093–5101. https://doi.org/10.1016/j.cub.2021.09.011

(2021).“), observed significant differences in spatial perception and sensorimotor responses between sighted and congenitally blind infants when engaging in tactile- and multimodal-guided reaching tasks. These findings strongly suggest that visual experience plays a crucial role in the development of multisensory integration and spatial calibration during early reaching behavior. However, the extent and implications of these developmental differences remain an open question, as compensatory mechanisms and neuroplasticity may allow for alternative strategies in the absence of vision. This developmental perspective underscores the importance of recognizing the diversity of sensorimotor profiles among individuals with visual conditions and the need for tailored assistive technologies that account for such variability in perceptual and motor strategies. In this light, our study does not aim to resolve these developmental questions rather acknowledges the complexity of the field. For us it suffices to note that the relationship between sensory experience and motor behavior is not clear-cut, and we handle our data within this broader context.

In recent years, advancements in assistive technologies have revolutionized the ability of individuals with visual conditions to navigate their environment, localize objects and complete tasks independently. Several assistive technologies have been developed to enhance the autonomy and well-being of individuals with visual conditions[12](#ref-CR12 “Velázquez, R. Wearable Assistive Devices for the Blind. In Wearable and Autonomous Biomedical Devices and Systems for Smart Environment Lecture Notes in Electrical Engineering (eds Lay-Ekuakille, A. & Mukhopadhyay, S. C.) (Springer, 2010). https://doi.org/10.1007/978-3-642-15687-8_17

.“),[13](#ref-CR13 “Manjari, K., Verma, M. & Singal, G. A survey on assistive technology for visually impaired. Internet Things https://doi.org/10.1016/j.iot.2020.100188

(2020).“),[14](https://www.nature.com/articles/s41598-025-21720-0#ref-CR14 “Seelman, K. D. et al. Quality-of-life technology for vision and hearing loss [Highlights of Recent Developments and Current Challenges in Technology]. IEEE Eng. Med. Biol. Mag. 27(2), 40–55. https://doi.org/10.1109/EMB.2007.907393

(2008).“). These assistive technologies leverage alternative sensory modalities, such as auditory[15](#ref-CR15 “Brown, D., Macpherson, T. & Ward, J. Seeing with sound? Exploring different characteristics of a visual-to-auditory sensory substitution device. Perception 40(9), 1120–1135. https://doi.org/10.1068/p6952

(2011).“),[16](#ref-CR16 “Levy-Tzedek, S., Hanassy, S., Abboud, S., Maidenbaum, S. & Amedi, A. Fast, accurate reaching movements with a visual-to-auditory sensory substitution device. Restor. Neurol. Neurosci. https://doi.org/10.3233/RNN-2012-110219

(2012).“),[17](https://www.nature.com/articles/s41598-025-21720-0#ref-CR17 “Ward, J. & Meijer, P. Visual experiences in the blind induced by an auditory sensory substitution device. Conscious. Cogn. 19(1), 492–500. https://doi.org/10.1016/j.concog.2009.10.006

(2010).“) or haptic[18](https://www.nature.com/articles/s41598-025-21720-0#ref-CR18 “Bach-Y-Rita, P., Collins, C. C., Saunders, F. A., White, B. & Scadden, L. Vision substitution by tactile image projection. Nature 221, 963–964. https://doi.org/10.1038/221963a0

(1969).“),[19](https://www.nature.com/articles/s41598-025-21720-0#ref-CR19 “Cancar, L., Díaz, A., Barrientos, A., Travieso, D. & Jacobs, D. M. Tactile-sight: A sensory substitution device based on distance-related vibrotactile flow. Int. J. Adv. Robot. Syst. https://doi.org/10.5772/5623

(2013).“) feedback, to compensate for the lack of visual input.

Auditory substitution is a popular method for compensating for lost vision because it leverages the brain’s cross-modal reorganization, particularly in people with visual conditions, allowing the dorsal visual stream to process auditory spatial information, effectively aiding in tasks that substitute vision with auditory cues[20](https://www.nature.com/articles/s41598-025-21720-0#ref-CR20 “Collignon, O., Lassonde, M., Lepore, F., Bastien, D. & Veraart, C. Functional cerebral reorganization for auditory spatial processing and auditory substitution of vision in early blind subjects. Cereb. Cortex 17(2), 457–465. https://doi.org/10.1093/cercor/bhj162

(2007).“). While auditory sensory substitution devices can enhance independence for people with limited vision, they may impede the primary auditory channel, masking essential sounds needed for localization[21](https://www.nature.com/articles/s41598-025-21720-0#ref-CR21 “Röder, B. et al. Improved auditory spatial tuning in blind humans. Nature 400, 162–166. https://doi.org/10.1038/22106

(1999).“),[22](https://www.nature.com/articles/s41598-025-21720-0#ref-CR22 “Brown, D. J. & Prolux, M. J. Audio-vision substitution for blind individuals: Addressing human information processing capacity limitations. EEE J. Sel. Top. Signal Process. 10(5), 924–931. https://doi.org/10.1109/JSTSP.2016.2543678

(2016).“). Tactile feedback devices, on the other hand, offer a promising alternative by using sensory augmentation to convey spatial information through touch, avoiding interference with critical auditory inputs[23](#ref-CR23 “Berning, M., Braun, F., Riedel, T. & Beigel, M. ProximityHat: a head-worn system for subtle sensory augmentation with tactile stimulation. In Proc. of the 2015 ACM International Symposium on Wearable Computers (ISWC ‘15), 31–38. https://doi.org/10.1145/2802083.280208

(2015).“),[24](#ref-CR24 “Duarte, B., McDaniel, T., Chowdhury, A., Gill, S. & Panchanathan, S. HaptWrap: Augmenting non-visual travel via visual-to-tactile mapping of objects in motion. In Proc. of the 2nd Workshop on Multimedia for Accessible Human Computer Interfaces (MAHCI ‘19) (pp. 17–24). Association for Computing Machinery, New York, NY, USA. https://doi.org/10.1145/3347319.3356835

(2019).“),[25](https://www.nature.com/articles/s41598-025-21720-0#ref-CR25 “Sait, U. et al. Design and development of an assistive device for the visually impaired. Proc. Comput. Sci. 167, 2244–2252. https://doi.org/10.1016/j.procs.2020.03.277

(2020).“). By leveraging the sense of touch, these devices enable users to interact with their surroundings more intuitively and safely, enhancing overall situational awareness. Tactile sensory substitution, like auditory substitution, effectively compensates for lost vision by engaging cross-modal reorganization in the brain, particularly in early visually conditioned people, allowing them to process tactile information through visual cortical areas and perform visual-like tasks using tactile cues[18](https://www.nature.com/articles/s41598-025-21720-0#ref-CR18 “Bach-Y-Rita, P., Collins, C. C., Saunders, F. A., White, B. & Scadden, L. Vision substitution by tactile image projection. Nature 221, 963–964. https://doi.org/10.1038/221963a0

(1969).“),[26](https://www.nature.com/articles/s41598-025-21720-0#ref-CR26 “Poirier, C., De Volder, A. G. & Scheiber, C. What neuroimaging tells us about sensory substitution. Neurosci. Biobehav. Rev. 31(7), 1064–1070. https://doi.org/10.1016/j.neubiorev.2007.05.010

(2007).“). These developments underscore the remarkable adaptability of the brain and the potential for assistive technologies to significantly enhance the independence and quality of life of those with visual conditions.

Building on the advancements in sensory substitution technologies, researchers have also explored vibrotactile feedback to aid individuals with visual conditions in navigating their surroundings. For example, the feelSpace naviBelt, which is a vibrotactile navigation aid worn around the waist that continuously signals the direction of magnetic north through subtle vibrations, enabling people with visual conditions to maintain orientation and navigate more confidently. It provides an intuitive, hands-free, and attention-sparing sense of direction, enhancing spatial awareness, emotional security, and independence in everyday mobility tasks[9](https://www.nature.com/articles/s41598-025-21720-0#ref-CR9 “Kärcher, S. M., Fenzlaff, S., Hartmann, D., Nagel, S. K. & König, P. Sensory augmentation for the blind. Front. Human Neurosci. https://doi.org/10.3389/fnhum.2012.00037

(2012).“),[27](https://www.nature.com/articles/s41598-025-21720-0#ref-CR27 “Brandebusemeyer, C., Luther, A. R., König, S. U., König, P. & Kärcher, S. M. Impact of a vibrotactile belt on emotionally challenging everyday situations of the blind. Sensors 21(21), 7384. https://doi.org/10.3390/s21217384

(2021).“),[35](https://www.nature.com/articles/s41598-025-21720-0#ref-CR35 “FeelSpace—naviGürtel. Retrieved July 30, 2025, from https://feelspace.de/

“). The feelSpace naviBelt has been further investigated, exploring how humans can learn new sensorimotor contingencies through long-term use of the belt, leading to integration of magnetic orientation cues, changes in navigation behavior, and profound alterations in spatial perception[8](https://www.nature.com/articles/s41598-025-21720-0#ref-CR8 “Kaspar, K., König, S., Schwandt, J. & König, P. The experience of new sensorimotor contingencies by sensory augmentation. Conscious. Cogn. 28, 47–63. https://doi.org/10.1016/j.concog.2014.06.006

(2014).“),[28](https://www.nature.com/articles/s41598-025-21720-0#ref-CR28 “König, S. U. et al. Learning new sensorimotor contingencies: Effects of long-term use of sensory augmentation on the brain and conscious perception. PLoS ONE https://doi.org/10.1371/journal.pone.0166647

(2016).“),[29](https://www.nature.com/articles/s41598-025-21720-0#ref-CR29 “Nagel, S. K., Carl, C., Kringe, T., Märtin, R. & König, P. Beyond sensory substitution—Learning the sixth sense. J. Neural Eng. https://doi.org/10.1088/1741-2560/2/4/R02

(2005).“). Its current iteration serves as an assistive device for individuals with visual conditions or blindness without interfering with the auditory sensory channel. Additionally, it can be integrated with a smartphone and a dedicated app, enabling its use as a wearable navigation system[35](https://www.nature.com/articles/s41598-025-21720-0#ref-CR35 “FeelSpace—naviGürtel. Retrieved July 30, 2025, from https://feelspace.de/

“). This technology highlights the potential for non-visual, non-auditory sensory feedback to improve autonomy and spatial orientation, enhancing the toolkit available for individuals with visual conditions.

While these studies have provided valuable insights into the potential of vibrotactile feedback for spatial guidance, the applications have been primarily focused on larger-scale navigation rather than the finer aspects of object manipulation, such as reaching and grasping. Such devices are still scarce, but some devices support people with limited vision in some way while grasping. For example, PalmSight30 is a device that provides tactile feedback via vibrating motors on the back of the user’s hand, utilizing visual information captured by a camera mounted on the palm. Further, GuideCopter[31](https://www.nature.com/articles/s41598-025-21720-0#ref-CR31 “Huppert, F., Hoelzl, G. & Kranz, M. GuideCopter—A precise drone-based haptic guidance interface for blind or visually impaired people. In: Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems (CHI ‘21). Association for Computing Machinery, New York, NY, USA, Article 218, 1–14. https://doi.org/10.1145/3411764.3445676

(2021).“) is a drone-based haptic guidance interface that employs a distinct form of haptic feedback to deliver fine-grained haptic cues and physically direct the user’s hand during hand-object localization tasks in unfamiliar environments. The drone can physically push or pull the user’s hand towards the target object, providing a more direct and intuitive way to guide their movements and interact with the environment.

While these existing devices offer promising ways to support grasping and hand-object localization, they present notable limitations, such as disrupting grasping or lacking practicality for daily use. The placement of tactile feedback on the hand could disrupt the user’s grasping ability. Additionally, a drone-based interface may not be suitable for everyday practical use, which is a critical consideration for individuals with visual conditions, as grasping is a particularly challenging task, and the device’s ease of use in daily life is essential. Moving forward, developing more user-friendly and effective tactile-based solutions will be crucial in overcoming these challenges and ensuring that individuals with visual conditions can navigate both their environments and fine motor tasks with greater ease and confidence.

In light of the limitations of existing tactile and sensory augmentation devices to support hand guidance and grasping, our study investigates an approach that seeks to overcome these challenges by developing a vibrotactile bracelet to assist individuals with visual conditions during object grasping[34](https://www.nature.com/articles/s41598-025-21720-0#ref-CR34 “BVA Presse—Blindheit. (n.d.). Augeninfo.de. Retrieved October 18, 2024, from https://www.augeninfo.de/offen/index.php?thema=407

“). The tactile bracelet aimed to leave the auditory channel unobstructed while providing precise information via vibrotactile signals on the wrist for hand guidance. By positioning the device on the wrist, the user’s hand remains unencumbered while providing sufficient spatial resolution to distinguish between the four vibration motors corresponding to the up, down, left, and right movement directions. Activating the respective motors guides the user’s hand to the approximate location of the target object, after which the participant can leverage their natural sensorimotor skills to complete the grasping task. This design approach focuses the device’s functionality on the localization and grasping of objects which is particularly challenging for individuals with visual conditions.

Here, based on the previous study in sighted blindfolded participants[32](https://www.nature.com/articles/s41598-025-21720-0#ref-CR32 “Powell, P. et al. Helping blind people grasp: Evaluating a tactile bracelet for remotely guiding grasping movements. Sensors 24(9), 2949. https://doi.org/10.3390/s24092949

(2024).“), we address the question of whether vibrotactile signals applied via a bracelet on the wrist are suitable to support hand movements and grasping in people with limited vision, compared to auditory signals. Thus, we conduct a comprehensive evaluation of a vibrotactile bracelet when used by participants with limited vision and whether it supports localization of target objects and grasping movements. This pilot study explores the device’s potential as a foundation for future assistive technologies. Ongoing research is testing a more autonomous version that replaces the experimenter with an object recognition AI for real-time feedback, marking a key step toward scalable, user-independent systems[33](https://www.nature.com/articles/s41598-025-21720-0#ref-CR33 “Furtak, M., Pätzold, F., Kietzmann, T., Kärcher, S.M. & König, P. Helping Blind People Grasp: Enhancing a Tactile Bracelet with an Automated Hand Navigation System. arXiv, Preprint https://doi.org/10.48550/arXiv.2504.16502

(2025)“).

In this study, we explore the effect of conditions (auditory and tactile feedback) as well as task familiarity on grasping performance in people with visual conditions using a vibrotactile bracelet. Specifically, we compare two conditions—auditory and tactile—with the goal of evaluating how participants adapt to the tactile system over time and whether tactile commands can be as good as auditory commands in assisting with grasping movements. Our primary measures of interest for the grasping task are response time (RT) of successful trials as a graded measure of participant engagement and system usability, as well as incorrect/correct responses (accuracy) as a binary measure to see whether participants understood the tactile signal correctly in the localization task. Based in the previous study by Powel et al.[32](https://www.nature.com/articles/s41598-025-21720-0#ref-CR32 “Powell, P. et al. Helping blind people grasp: Evaluating a tactile bracelet for remotely guiding grasping movements. Sensors 24(9), 2949. https://doi.org/10.3390/s24092949

(2024).“), RT continues to reflect learning and adaptation, hence we chose to analyze only successful trials to capture meaningful differences in participant performance in the grasping task.

Specifically in our research, we propose three hypotheses: 1) the Modality Hypothesis, which states that participants’ RTs differs significantly between two feedback modalities, the auditory and tactile conditions. This hypothesis tests whether tactile feedback deviates from the auditory feedback; 2) the Learning Hypothesis, which states that participants RTs improve significantly over time, suggesting that task familiarity across blocks facilitates learning and enhances performance in both feedback modalities; and 3) the Position Hypothesis, which suggests that the spatial position of the target object will significantly influence RTs across both feedback modalities, reflecting the variability of object locations in real-life grasping tasks.

The aim of our study is to investigate whether tactile feedback can support efficient performance comparable to auditory feedback, whether repeated exposure leads to performance improvements over time, and whether object location systematically affects response times across feedback conditions, to provide a viable alternative to the auditory feedback.

Methods

Participants

To maximize the recruitment of participants with visual conditions, we were fortunate to leverage the database provided by the company “feelSpace”. We reached out to individuals who were in contact with feelSpace before and had permitted further contact. Moreover, we contacted the ‘Blindenverband Niedersachsen’ (Blind Association of Lower-Saxony) as well as other blind associations close to Osnabrück via email, requesting the possibility of including our recruitment message in their email newsletter to enhance outreach. Afterward, we received several inquiries from smaller organizations. They expressed interest in the possibility of us conducting testing sessions in their respective cities. This interest stemmed from the fact that many visually conditioned people are not at ease with traveling long distances to Osnabrück for participation. Furthermore, we had the opportunity to attend a exhibition focused on people with visual limitations in Hannover alongside feelSpace. During the event, we engaged with attendees on-site and successfully recruited participants spontaneously.

We successfully recruited a total of 42 participants with visual conditions for our study, comprising 21 females and 21 males, this is the official medical categorization[34](https://www.nature.com/articles/s41598-025-21720-0#ref-CR34 “BVA Presse—Blindheit. (n.d.). Augeninfo.de. Retrieved October 18, 2024, from https://www.augeninfo.de/offen/index.php?thema=407

“). Participants who experienced persistent migraines, headaches, tremors, uncorrectable speech or hearing difficulties, epilepsy, memory impairments, pregnancy, or were under the age of 18 were excluded from the study at the outset. Due to an experimental error, one participant’s data was excluded from the analysis, resulting in a final sample size of 41 usable participants.

Of the remaining 41 participants, 17 were congenitally blind, while 24 experienced acquired blindness later in life. Among the participants, 34 were completely blind, with 20 of them able to perceive light and/or detect changes in lighting conditions. Three participants were 98% visually impaired, all of whom could perceive light and/or lighting changes. Two participants were 95% visually impaired, both of whom also perceived light and/or lighting changes. One participant was 90% visually impaired and was capable of perceiving light and/or changes in lighting, while one participant was 80% visually impaired and did not provide information regarding their ability to perceive light or lighting changes.

Technical setup

The bracelet, shown in Fig. 1B, employed in our study is a prototype derived from the feelSpace naviBelt[9](https://www.nature.com/articles/s41598-025-21720-0#ref-CR9 “Kärcher, S. M., Fenzlaff, S., Hartmann, D., Nagel, S. K. & König, P. Sensory augmentation for the blind. Front. Human Neurosci. https://doi.org/10.3389/fnhum.2012.00037

(2012).“) and the technical aspects are documented in detail in a previous publication[34](https://www.nature.com/articles/s41598-025-21720-0#ref-CR34 “BVA Presse—Blindheit. (n.d.). Augeninfo.de. Retrieved October 18, 2024, from https://www.augeninfo.de/offen/index.php?thema=407

“). The bracelet consists of four vibration motors, which were also used for the feelSpace naviBelt, indicating each of the four directions (top, bottom, left, and right), providing vibrotactile input at 0°, 90°, 180°, and 270° around the participant’s wrist, and is worn on the dominant hand. The vibration feedback in our setup was generated using cylindrical eccentric rotating mass (ERM) motors (Precision Microdrives, Model 306-10H). These coreless, precious metal brush DC motors have a 7 mm body diameter and 24.5 mm length, operate at a rated voltage of 3 V, and produce a typical vibration amplitude of 1.84 G at 13,800 rpm with a nominal operating current of 50 mA. The ERM principle produces vibration by rotating an unbalanced mass around the motor shaft, converting the rotational motion into a haptic stimulus suitable for tactile feedback applications. The vibration characteristics can be highly adapted; in this study, the default intensity was set to 100 on a scale from 0 to 100.

Fig. 1

(A) Experimental setup: This figure presents the experimental setup, which includes nine plastic fruits on a shelf. (B) Vibration bracelet: This figure illustrates the bracelet’s design and how the participants wear it.

In the current version, the bracelet uses a fixed mapping on allocentric coordinates. That is, when a rightward movement is needed and the bracelet is worn on the right arm, the ulna vibration motor is switched on. At this stage of development the tactile information does not adjust to different hand orientations. Further, this is done independent of any pronation or supination of the hand. Consequently, the ulna vibration motor, that in a palm down hand gesture is on the lateral/right side, might point medially when the hand in an extreme supination position.

The motors are attached to a flexible strap that can be adjusted to fit the participants’ wrist sizes. The bracelet was controlled remotely by the experimenter via a laptop, which transmitted commands through a control box attached to the participant’s upper arm. The experimenter only had access to a webcam feed showing the participant’s hands and the target objects on the shelf. Auditory guidance was provided through the laptop’s speakers, where a computer-generated voice repeated the directional command (e.g. “right”, “left”, “up”, “down”, and “grasp”) until the experimenter pressed a button to change the command. Typically, one or two directional commands were followed by a grasping command, which was likewise repeated until the participant grasped and slightly lifted the target object. The experimenter manually terminated each trial upon task completion and documented its success, while the program simultaneously measured and recorded the participant’s response time. The response time includes all hand movement steps, moving in the indicated directions as well as the grasping movement.

In the tactile condition, the same procedure was used, but instead of auditory cues, the bracelet delivered vibrotactile stimuli, the directional vibrations were continuous, lasting as long as the participant needed to move their arm in this direction. The vibration for the grasping command was a pulsation, meaning the upper vibration module would pulsate to give a different sensory sensation than the directional cues. Each pulse was active for 150 ms. This means the signal stayed “on” briefly before switching off again. The total time from the start of one pulse to the start of the next was 500 ms. Since the “on” time was 150 ms, each pulse was followed by 350 ms in the “off” state. Each series contained exactly nine such pulses. At a 500-ms period per pulse, the complete set of pulses lasted 9 × 500 ms = 4500 ms. The time from the start of one pulse series to the start of the next is 5000 ms. This means there was a 500-ms pause after the last pulse before the next series would begin. The pulsation series would repeat as long as the participant needed to grasp the target object and lift it a few centimeters, after which the trial was terminated by the experimenter.

The experimenter only used the arrow keys on the keyboard to indicate the direction command, the “g” button for giving the grasping command, as well as the “s” or “f” button to record if the trial was successful or not. Each button only needed to be pressed once, and the stimuli (auditory and tactile) would persist until another button is pressed.

Participants were seated in front of a three-tiered shelf on which nine plastic fruits were positioned as target objects, each within comfortable grasping distance. The experimental setup is illustrated in Fig. 1A*.*

The camera, an Insta 360 One R with a 1-inch wide-angle lens, is positioned on a standard bicycle helmet to capture a clear view of the region the participant faces. This enabled the experimenter to use this image to provide cues for different directions and the grasping command.

Procedures

Upon arrival, participants were greeted, briefed about the experiment and task to be completed, and given consent forms and a medical questionnaire. These explanations were given verbally or in audio files, and participants may provide verbal answers and consent if preferred. The participants were informed that they could ask questions at any point and withdraw from the study without any legal implications.

After reviewing, understanding, and signing all necessary documents, participants were guided to a shelf containing artificial fruits, where the experimental tasks would occur. The participants were seated in front of the shelf within an individually adjusted distance considering varying arm’s length. The 9 plastic fruits were placed on the shelf at horizontal and vertical distances of ~ 30 cm and at a depth (measured from the front of the shelf) of about 10 cm. On average the distance of participant to the front of the shelf was about 60 cm and allowed comfortable grasping of the fruits with small movements of the upper body (Fig. 1A). The main experiment involved two conditions: an auditory condition, where verbal directions from a laptop guided hand movement, and a tactile condition, where the bracelet’s vibrations provided directional cues. In consideration of the visually conditioned participants, we chose to begin with the more familiar setup using auditory commands, aiming to enhance their subjective sense of security and support the overall success of the experiment. Participants were instructed to move their arms according to the auditory stimulus. They performed 9 auditory grasping trials across three blocks in this condition (27 in total), with movements closely monitored by a helmet-mounted camera for accuracy and guidance.

After a short break, participants were assisted in putting on the bracelet. To assess their ability to differentiate the signals, each motor vibrated once in a sequence (up, right, down, left), the vibration was active for 500 ms before switching off. This served as a pretest to ensure participants were comfortable wearing the bracelet, to give them the opportunity to familiarize themselves with the vibration sensations, and to verify the bracelet was positioned correctly for each participant, as wrist sizes differ quite a bit. This procedure was repeated until the participants felt comfortable and confident in differentiating the signals from the different directions; on average, participants required three rounds. This also allowed us to adjust the motor positions according to the participants’ needs, ensuring they could differentiate between the signals before starting the actual tactile part of the experiment: the localization task.

Following this, the participants completed the localization task with the tactile bracelet to train themselves with directional vibration signals, ensuring they could accurately identify and respond to tactile cues. After participants completed the localization task, the tactile grasping task would start. Each participant performed 9 grasping trials across three blocks (27 in total) in the tactile condition as before in the auditory condition. After completing the tasks, participants completed a feedback questionnaire to assess their experience and the bracelet’s effectiveness, offering insights into its potential utility for those with visual conditions.

Localization task

We conducted a localization task to test whether participants could accurately identify and respond to tactile cues. It also served as training for using the bracelet in the later grasping task. Once the participants were comfortable identifying the signals and understanding how the bracelet worked, the localization task would begin. The bracelet vibrated once, again for 500 ms, at randomly selected locations, and participants had to loudly indicate the direction of the vibration. This task did not involve hand movements, participants were instructed to sit in a comfortable position and resting their arm either on their leg or the armrest of the chair with the bracelet not touching any of it, to minimize the possibility of malperception of the vibrational signals (Fig. 2). 16 trials comprised one block, and the localization task consisted of three such blocks, summing up to 48 trials in total. This localization task aims to evaluate participants’ ability to identify the location of the tactile stimuli while also serving as a training exercise to familiarize them with the vibration sensation and directional commands.

Fig. 2

Conceptual representation of the localization task.

If participants exhibited discomfort in their responses or if the mean accuracy of the localization task was below 90% after the three blocks, additional trials—typically two supplementary blocks—were administered. Accuracy was then calculated based on performance in the final three blocks. In 14 cases, when the accuracy did not reach 90% even after additional training, we proceeded with the experiment, nonetheless. The participants showed improvement in the task, which we deemed significant progress.

Grasping task

In the main experiment, containing the auditory as well as the tactile condition, participants grasped and slightly lifted nine plastic fruits placed on the shelf in front of them in a computer-generated random sequence before putting them down again (Fig. 1A). This formed one block of 9 trials, each trial started with giving the first command and lasted until the participant grasped the fruit, we conducted three blocks in each condition summing up to a total of 27 trials per condition and 54 trials in total. The fruits had prototypical forms and were within reach. Participants were instructed to position their hands at the central shelf’s designated starting point (see Fig. 6B) and return to this position after each grasping action. Participants received a demonstration of the operational protocols for auditory and tactile conditions to ensure a comprehensive understanding. Each participant received at least three examples, facilitating familiarity and clarity in task execution before the experimental trials, if this was not enough, they received another round of three demonstrations.

In both conditions, the experimenter observed the scene via the webcam and guided the participants using verbal or tactile commands. Each trial was manually initiated and terminated by the experimenter using a button press. After each trial, the experimenter recorded whether the participant had successfully grasped the target object or if the trial was unsuccessful due to errors made by either the participant or the experimenter (Fig. 3).

Fig. 3

Conceptual representation of the grasping task.

In the auditory condition, participants received verbal instructions from a laptop, which included commands like “left”, “right”, “up”, and “down” to guide their hand movements. They were also instructed to perform a grasping movement in response to the word “grasp”. Whereas in the tactile condition, participants received vibrotactile instructions indicating the “left”, “right”, “up”, and “down” directions through continuous vibrations of the respected motor and a short, consecutive vibration from the upper motor to indicate when to grasp. We utilized the laptop’s arrow keys to impart an impulse for the desired direction for both conditions and the “g” key for indicating the grasping command.

Questionnaire

After completing the experiment, participants were asked to complete a questionnaire of eleven questions, each rated on a five-point Likert scale, where ‘1’ indicates total disagreement and ‘5’ indicates total agreement. At the end of the questionnaire, participants were also given the opportunity to provide additional feedback through open-ended questions. The post-experiment feedback provided valuable insights into the perceived utility of the bracelet among the target demographic. To analyze the responses, we conducted a factor analysis (FA) to identify underlying latent groupings.

To further analyze the qualitative data, we developed a categorization system. Initially, we transcribed the open-ended responses into English. We then created a category system to quantify relevant aspects of the feedback and identify key themes in participants’ experiences. This system was modeled after the framework established by Kaspar et al.[8](https://www.nature.com/articles/s41598-025-21720-0#ref-CR8 “Kaspar, K., König, S., Schwandt, J. & König, P. The experience of new sensorimotor contingencies by sensory augmentation. Conscious. Cogn. 28, 47–63. https://doi.org/10.1016/j.concog.2014.06.006

(2014).“). The categorization process was supported by ChatGPT 3.5, which generated preliminary categories based on our data that aligned with Kaspar et al.’s approach. We provided a framework outlining the desired structure of the system and presented our translated statements. The model was then tasked with developing a category system aligned with this framework and organizing our statements accordingly.

We refined the initial ChatGPT generated categories, addressing overlaps and redundancies, ultimately resulting in five Level-1 Categories, each encompassing various Level-2 and Level-3 categories. Definitions were then formulated for each subcategory, and team members independently sorted the participants’ statements into the established categories. Additionally, we presented ChatGPT with our finalized categories and requested that it re-sort our statements according to this updated categorization system. Upon comparison, we found a high degree of agreement in our categorizations, with only minor discrepancies in a few statements, resulting in our final Category System described further below.

Results

Localization task

We conducted a localization task to test whether the 41 participants could correctly identify the tactile input from the bracelet. On average, participants achieved an accuracy of 90.6% in localizing the tactile input on the wrist (Fig. 4A). This is slightly above the target accuracy threshold of 90% across the group. However, the spread of accuracy achieved by different participants varied considerably. Out of all the participants, 13 reached a perfect score of 100%, and an additional 15 scored above 90%. Nine participants missed out target accuracy by less than 10%, and five participants had much lower scores. This indicates that the design of the bracelet is working well in many cases but has to be adapted to the individual needs in many other cases.

Fig. 4

(A) Mean accuracy in the localization task by participant. The participants are sorted according to the accuracy reached. The red dashed line represents the mean accuracy across all participants, providing a reference point for assessing individual performance. (B) Confusion matrix of vibration signals gives the probability of interpretations for each signal. Correct interpretations are on the main diagonal; confusion of neighboring tactile elements is on the off-diagonal and upper left and lower right corner.

In Fig. 4B we analyzed how participants interpreted each directional command using the confusion matrix to evaluate the effectiveness of the tactile signals. The confusion matrix summarizes the interpretation of all signals averaged over participants. In the figure, y-axis shows the directional command and x-axis represents the response given by participants. The entries on the main-diagonal, marked with dark grey, shows the correct response, and any off-diagonal entry indicates that a certain vibration element was mistaken as another. We also added cases that participants didn’t give a response to a specific command, which shown in the last column. For example, the direction command “left” was misidentified as “up” in 5.5% of cases, as “right” in 0.6%, and as “down” in 0.8%. It was correctly identified in 87.6% of cases, while no response was given in 5.5% of cases. The accuracy differs only slightly between different directional commands, with down being the most reliable and left and up the most error prone. The errors often occur in neighboring tactile elements, notably up being interpreted as right, right as down, and left as up. Overall, these findings suggest that the separation of tactile elements works well, but the interpretation of neighboring tactile elements has room for improvement.

Grasping task

To evaluate the effectiveness of tactile feedback compared to auditory feedback in guiding grasping movements, we examined the mean response times over participants across both tactile and auditory conditions from the start of the first signal until grasping the object by only including successful trials. The results showed that the mean trial times for the auditory condition (M = 4198 ms; SD = 1691 ms) were faster than those in the tactile condition (M = 4784 ms; SD = 2255 ms), resulting in an average advantage for the auditory condition of 586 ms (Fig. 5A). Nevertheless, 15 of 41 participants performed faster grasping movements with the tactile guidance than with auditory guidance. Thus, while the auditory condition generally yielded superior results, our findings suggest that tactile feedback also served as an effective alternative for grasping tasks for people with visual conditions. Next, we analyzed the mean response time across the course of the experiment. It is highest in the first block and then decreases in subsequent blocks. Compared to the change from the first to the second block, the change from the second to the third block is much smaller.

Fig. 5

(A) Violin plot shows the distribution of response times for all individual grasping movements under auditory and tactile feedback conditions. The plot includes all data points across the entire dataset (not averaged per participant). The box within each violin indicates the interquartile range (IQR), the central line marks the median response time, and the red point indicates the overall mean. (B) Violin plot shows the distribution of all individual response times across the three experimental blocks (reflecting task familiarity). Again, the plot includes all data points across the dataset. The boxes represent the IQR, the central line shows the median, and the red point indicates the mean.

To evaluate how feedback modality (tactile and auditory conditions) and task familiarity (blocks 1, 2, and 3) influence grasping performance, we applied a linear mixed model. This model accounted for fixed effects of feedback modality and task familiarity, with random intercepts and slopes to account for individual differences. The model explained 4% of the variance through fixed effects alone (marginal R2 = 0.04) and 56% when including random effects (conditional R2 = 0.56), suggesting that individual differences contributed substantially to the total explained variance. The analysis revealed a significant main effect of feedback modality (p < 0.005), with auditory guidance being faster than tactile guidance. Moreover, task familiarity also had a significant effect on performance. Grasp times improved across blocks, decreasing by 504 ms (p < 0.001) in Block 2 and by 624 ms (p < 0.001) in Block 3, indicating that participants became faster with practice (Fig. 5B). Although we did not include an interaction term in the final model, our preliminary analyses revealed no significant interaction between feedback modality and task familiarity. This suggests that the effect of task familiarity on grasping performance was independent of the type of feedback provided. This result also aligns with our Learning Hypothesis that learning to interpret the signal was equally fast for both conditions. Overall, while auditory feedback proved more effective in improving grasping performance, participants became faster in both conditions as they gained experience, suggesting that tactile feedback remains a viable alternative.

To better understand the role of spatial layout in grasping performance, we conducted an additional analysis based on the Position Hypothesis, which proposes that grasping response times are affected by the spatial location of the object, particularly its distance from the central starting position (position 5). The hypothesis assumes that increasing spatial distance increases the variability in RTs. Target positions were grouped into two categories: corner positions (positions 1, 3, 7, and 9), which required three motor commands (two directional and one grasping), and edge positions (positions 2, 4, 6, and 8), which required only two commands (one directional and one grasping). See Fig. 6B for an illustration of the shelf layout. A paired-samples t-test revealed that participants responded significantly faster to edge positions (M = 3797.35 ms, SD = 1056.10) than to corner positions (M = 5207.48 ms, SD = 1640.22), t(40) = 13.52, p < 0.001, 95% CI [1199.30, 1620.95], with a mean difference of 1410.12 ms (Fig. 6A). These results suggest that grasping performance is significantly influenced by both the spatial distance from the starting position and the number of required motor commands.

Fig. 6

(A) Mean response time for corner vs edge positions. The bar graph displays mean response times (