Cognitive Evaluation Of The Fluxtactile Communication

José Carlos da Cunha1,2, Percy Nohama2
1 Department of Computer Engineering, Positivo University, Brazil
2 Post-Graduation Program in Electrical Engineering and Computer Science, Federal Technological University of Paraná (UTFPR), Brazil


Fluxtactile stimulation has been studied to promote alternative communication for people with visual and hearing impairments. An instrumentation was developed to investigate the cognitive responses of the tactile perception when submitted to CO2 jets. Twenty five males subjects, without physical and cognitive disabilities, were submitted to the developed alternative system and their cognitive response were evaluated over a set of 19 characters that were transferred to the abdominal region of them in four plotting speeds. The results of this study indicate that: (1) 40 mm/s-plotting speed produces the best correct recognition rate among the tested values; (2) characters shape influences on the hit rate, varying from 10% up to 98%. These results suggest that Meisner Corpuscules are the main responsible by the fluxtactile stimulation since 40 mm/s-plotting speed produce a 1.33 Hz relative frequency stimulation on the abdominal region, where the receptive field has the size of 30 ± 5 mm. These data may guide designers in their new projects of alternative communication, assistive technology and augmented reality devices.


Electrical and mechanical tactile stimulation have been used as an alternative channel of communication to promote an interface between the world and people, especially those with visual and/or hearing impairing (Kaczmarek et al., 2006). Mechanotactile stimulation through air jets has been used in neurology, psychology and psychophysics in order to investigate and diagnose neurological functions (Hamalainen et al., 1985; Hashimoto et al., 2000). To promote an alternative way of communication, alerts and/or augmented reality in car driving, machine operation, games, some researchers have been used air jet stimulation on devices which converts visual or hearing information into a controlled flow of gas applied to the skin (Gwilliam, et al., 2012; Rantala and Raisamo, 2011; Bianchi et al., 2011; Cunha et al., 2010, Furuya et al., 2009).

The skin contains several types of sensory receptors classified as mechanoreceptors. Meissner and Pacinian corpuscles are the most important skin sensory receptors responsible for the touch (Lent, 2010). Each mechanoreceptor responds to skin deformation and motion in a different way (Bear et al., 2006).

Cunha et al. (2010) developed a fluxtactile stimulation system yielding CO2 jet on the skin in order to promote and study a novel alternative for communication. The device is a 2-D plotter that can draw alphabetic characters on the skin through a stimulation needle.  

The psychophysics responses of this stimulation were investigated and discussed in Cunha & Nohama (2012), where the threshold perception force produced by the gas jet stimulation as a function of the pressure and flow, and the plotting speed perception influence were investigated. This threshold was otained with 10 volunteers and showed a mean value of 0.6 ± 0.0019 mN for mean pressure of 30 kPa and flow of 4 L/min. Plotting speeds of 20, 40, 67 and 100 mm/s were assessed and the best tactile perception occurred for 40 mm/s.


The purpose of the experimental study introduced in this paper was to answer the following research questions:

  1. Does the fluxtactile stimulation elicit a profitable statistical correct recognition rate over a set of 19 characters as a function of the cognitive aspects?
  2. Do the character’s shapes and plotting speed influence the hit rate of each character?



The population of interest in this study was persons without visual, hearing or cognitive impaired. Therefore, were recruited 25 male able-bodied subjects to conduct this study, ranged in age from 18 to 35 years. Were excluded from this study females in function of the tactile sensitive variation during the menstrual period. 


fillerFigure 1: general representation of the fluxtactile system. Adopted from Cunha et al. (2010).

The fluxtactile stimulation system developed for this study (Cunha et al., 2010) is composed by a pneumatic structure and an electronic circuit, which is responsible for the generation and transference of stimulatory patterns to the subject under investigation by means of a 2D tactile plotter. It allows draw characters using a thin CO2 jet applied to the skin through a 1 mm-internal diameter needle, which is connected to the pneumatic block and positioned at the stimulating head of the fluxtactile plotter. Figure 1 shows a general representation of the system developed to this study.

The characters generated in the system, whose shape and size are illustrated in Figure 2, may be drawn at different speeds through the developed plotter-computer interface. Pressure and flow of the gas were adjusted to promote a force of 1.8 mN, equivalent to three times the sensation threshold, ensuring the perception of the gas jet by all the volunteers. According to Cunha & Nohama (2012), the force produced by the gas jet on the skin remains constant for a distance between the needle tip and the skin surface, ranging between 3 and 25 mm.

Experimental Protocol

FillerFigure 2: The shape, size and “drawing” orientation to the plotting characters used in the validation tests.

The experimental protocol was performed as follows: (1) the fluxtactile plotter was placed at the abdominal region of the volunteer in standing position, (2) for each test, the plotter had its height adjusted according to the height of the volunteer so that the excited region remained the same, (3) the distance between the injection needle and the skin surface was controlled to provide an average distance of 10 mm with a minimum of 5 mm and a maximum of 15 mm, depending on the curvature of the abdominal surface (4) the tests were conducted at laboratory temperature (22 °C), inside an environment without airflow on the position in which the equipment was installed, and with the gas jet at a temperature of 24 °C, (5) before testing, each subject underwent a 30-minute training where he learned to recognize each character plotted in his abdomen, (6) after the training period, each volunteer was submitted to a blind test in which they were asked about the eleven random characters plotted on his skin, from a total of nineteen ones as shown in Figure 1, (7) four sets of tests with plotting speeds of 20, 40, 67 and 100 mm/s were applied, (8) for each plotting speed, tests were applied on 15 subjects, totaling 60 tests. (9) the results were recorded on a worksheet for further statistical analysis.

Data Analysis

Were analyzed the hits rate as a function of the plotting speed and determined the influence of the shape and plotting speed on these results. An ANOVA was applied over these results, considering the four plotting apeeds set. 


fillerFigure 3: Thecorrect recognition rate for all characters for the 20, 40, 67 and 100 mm/s plotting speeds.

Figure 3 illustrates the correct recognition rate for all the 19 characters tested as a function of the four speeds of plotting. It can be observed that recognition rate presented great variations for different alphabetic characters as well as for different plotting speeds.    

fillerFigure 4: :  Mean correct recognition rate and its standard deviation for all characters on a crescent order.

Among the characters, the most significant variations in the correct recognition rate were observed on S, G and M characters, reaching 63%, 47% and 46%, respectively, which resulted in the highest SD (25%, 21% and 18%, respectively), while the smaller variations were observed on I, J and U characters, with rates of 10%, 14% and 14% and SD of 4, 6 and 6%, respectively. These results are illustrated in Figure 4.      

fillerFigure 5: The mean and SD hit rate for the plotting speeds of 20, 40, 67 and 100 mm/s.

Figure 5 shows the mean and SD correct recognition rate for the four evaluated plotting speeds. A global hitting rate of 59.6 ±13,5 % was obtained. The best hit rate was observed with the speed of 40 mm/s, followed by the speed of 67 mm/s, with 68 ±17 % and 60 ±15 %, respectively, and the worst results were observed with the speeds of 20 and 100 mm/s with rates of 54 ±20 % and 56 ±20 %, respectively.


The obtained results for correct recognition rate did not show statistical significant variation as a function of the plotting speeds, although the best rates were observed at 40 mm/s. Considering that the tests were applied at the abdominal region, where the receptive fields have about 30 ±5 mm, the stimulation needle displacement over the skin at 40 mm/s produces a relative stimulation frequency of 1.33 Hz. This indicates that Merkel discs were activated, evoking a tactile response. According to Bear et al. (2008) and Lent (2010), Merkel discs responds between 0.3 and 3.0 Hz, constituting the most specialized mechanoreceptors for detection of static or low-frequency pressure.

Regarding the cognitive response, although all the subjects have been received 30-minute training, it was verified a great variation in the characters hit rate, with a great variation in the SD, as observed in Figure 4. The higher SD were observed with the characters S, G, M, L, P, E and V with 25%, 21%, 18%, 17%, 17%, 14% and 14%, respectively.

Another recognition problem is related to the shape and direction of plotting of some specific characters, which hold similarity to each other. These are the case of D and O, F and E, J and U, M and N, and U and V. This indicates that graphic design of characters may have an impact on speed and accuracy of perception and recognition. Although this subject was not formally studied, personal reports of some of the volunteers indicate that the character’s size and plotting speeds must be set to the best results.


The experiments performed on twenty-five subjects showed that the developed instrumentation is adequate to evaluate the cognitive response to fluxtactile stimulation.

The results of this experimental study indicate the viability of the fluxtactile stimulation method in alternative communication applications and so for deaf and/or blind people, beyond the operation of machines and vehicles, devices for guided surgery and games.

Further tests with deaf, blind or deafblind people will be performed in the future, in order to investigate its real efficacy with these specific users. Its use in biofeedback for artificial limbs and for augmented information systems such as in car driving by handicapped, and perception in dark environments are other possibilities that may be investigated.


Bear, M. F., Connors, B. W., Paradiso, M. A. (2006).  Neuroscience: exploring the brain.   Lippincott Williams & Wilkins. Third Edition.

Bianchi, M.; Gwilliam, J.C.;  Degirmenci, A. and Okamura, A. M. (2011).  Characterization of an Air Jet Haptic Lump Display. in Proc. IEEE Engineering in Medicine and Biology Society, 2011, pp. 3467-3470.

Cunha, J. C.; Bordignon, L. A.; Nohama, P (2010). TactileCommunication Using a CO2 Flux Stimulation for Blind or Deafblind People. In Proc. 32nd Annual International IEEE EMBS Conference, August 31 - September 4, 2010, Buenos Aires Sheraton Hotel, Buenos Aires, Argentina. 2010.

Cunha, J. C. and Nohama, P. (2012). Estudo da RespostaPsicofísica à Estimulação Fluxotátil em Aplicações de Comunicação Alternativa. In Proc. XXIII Congresso Brasileiro em Engenharia Biomédica – XXIII CBEB.

Furuya, T.; Yanagisawa, Y.; Tamesue, T.; Itoh, K. (2009)Transmission of Acoustic Information of Percussion Instruments through Tactile Sensation Using Air-Jet Stimulation for Hearing Impaired Person. In: Universal Access in Human-Computer Interaction. Addressing Diversity, V. 5614, pp. 48 – 57. Springer Berlin / Heidelberg.

Gwilliam, J. C.; Degirmenci, A.; Bianchi, M. and Okamura,A. M. (2012). Design and control of an air-jet lump display. In Proc. IEEE Haptics Symposium, 2012, pp. 45-49.

Hamalainen, H.A.; Warren, S.; Gardner, E.P., 1985. Differential sensitivity to airpuffs on human hairy and glabrous skin. Somatosens Res. 2 (4): 281-302.

Hashimoto, I.; Yoshikawa, K.; Kimura, T. (2000). Somatosensory evoked potentials to a threshold air-puff can predict stimulus detection in human subjects. Neuroscience Letters. Mar 24; 282 (3):181-4.

Kaczmarek, K. A., 2000. Electrotactile adaptation on the abdomen: Preliminary results, IEEE Transactions on Rehabilitation Engineering 8, 499-505.

Lent, R. (2010). Cem Bilhões de Neurônios – Conceitos Fundamentais de Neurociência, Rio de Janeiro: 2 ed. Atheneu.

Rantala, J. &  Raisamo, J. (2011). Tactile Sensing &Feedback. In: HUI 2011. Tampere Unit for Computer-Human Interaction. School of Information Sciences University of Tampere, Finland. 2011.