Technology and Engineering HuMOVE: A Low-invasive Wearable Monitoring Platform in Sexual Medicine Gastone Ciuti, Matteo Nardi, Pietro Valdastri, Arianna Menciassi, Ciro Basile Fasolo, and Paolo Dario OBJECTIVE
MATERIALS AND METHODS
RESULTS
CONCLUSION
To investigate an accelerometer-based wearable system, named Human Movement (HuMOVE) platform, designed to enable quantitative and continuous measurement of sexual performance with minimal invasiveness and inconvenience for users. Design, implementation, and development of HuMOVE, a wearable platform equipped with an accelerometer sensor for monitoring inertial parameters for sexual performance assessment and diagnosis, were performed. The system enables quantitative measurement of movement parameters during sexual intercourse, meeting the requirements of wearability, data storage, sampling rate, and interfacing methods, which are fundamental for human sexual intercourse performance analysis. HuMOVE was validated through characterization using a controlled experimental test bench and evaluated in a human model during simulated sexual intercourse conditions. HuMOVE demonstrated to be a robust and quantitative monitoring platform and a reliable candidate for sexual performance evaluation and diagnosis. Characterization analysis on the controlled experimental test bench demonstrated an accurate correlation between the HuMOVE system and data from a reference displacement sensor. Experimental tests in the human model during simulated intercourse conditions confirmed the accuracy of the sexual performance evaluation platform and the effectiveness of the selected and derived parameters. The obtained outcomes also established the project expectations in terms of usability and comfort, evidenced by the questionnaires that highlighted the low invasiveness and acceptance of the device. To the best of our knowledge, HuMOVE platform is the first device for human sexual performance analysis compatible with sexual intercourse; the system has the potential to be a helpful tool for physicians to accurately classify sexual disorders, such as premature or delayed ejaculation. UROLOGY 84: 976e981, 2014. 2014 Elsevier Inc.
O
ver the past few decades, the increased level of awareness within health and physical fitness communities has created an emerging need for smart applications and comfortable wearable devices that sense, classify, and provide information about users’ activities of daily living in a pervasive fashion.1-6 In particular, physical motion, which is a fundamental aspect for categorizing a person’s current health and activity, can be detected and recorded by using wearable accelerometer-based systems.7,8 Their recent impact is mainly because of their small size, low invasiveness, intuitiveness, relatively low cost, and easy integration with existing platforms for sensor networks. They are also able to effectively provide indicative features of human
Financial Disclosure: The authors declare that they have no relevant financial interests. From the The BioRobotics Institute, Scuola Superiore Sant’Anna, Pisa, Italy; the STORM Lab, Department of Mechanical Engineering, Vanderbilt University, Nashville, TN; and the Department of Clinical and Experimental Medicine, University of Pisa, Pisa, Italy Reprint requests: Gastone Ciuti, Ph.D., The BioRobotics Institute, Scuola Superiore Sant’Anna, Viale Rinaldo Piaggio, 34, Pontedera 56025, Pisa, Italy. E-mail: g.ciuti@ sssup.it Submitted: February 11, 2014, accepted (with revisions): June 27, 2014
976
ª 2014 Elsevier Inc. All Rights Reserved
movement corresponding to specific parameters of motion that are important to assess, for instance, static posture characteristics.9 An important field that may benefit from the use of a wearable monitoring inertial device (WMID) is sexual medicine, because of its consistent impact and influence on human life quality and the lack of quantitative and low-invasive monitoring systems for objective parameter analysis during sexual intercourse (SI). Nowadays, basic and/or even invasive methodologies often not compatible with SI are used for diagnosing sexual disorders (eg, premature ejaculation), such as simple stopwatches (mainly used for assessing the intravaginal ejaculation latency time parameter [IELT]), invasive button-counter for calculating SI penetrations, and personal evaluation forms about the satisfaction level of the partner.10-14 Basing on the aforementioned premises, a wearable inertial monitoring platform based on accelerometer sensing, suggested and named Human Movement (HuMOVE) by the author C.B.F., was developed by the authors with the aim to sense, record, and derive inertial movement parameters for evaluating human sexual performance for diagnostic purposes. Although a http://dx.doi.org/10.1016/j.urology.2014.06.040 0090-4295/14
Figure 1. (A) HuMOVE: an accelerometer-based analysis solution for human sexual performance evaluation. (B) System architecture of the overall system composed of a WMID, an MBT, and a remote station. (C) Wearable monitoring inertial device positioned on the human body and working principle and (D) software workflow: once the system is activated and initialized in the “waiting mode” from the “sleeping mode” (phase 1), the user has to approach the MBT with the different sides (phase 2 for recording mode and phase 3 for sleeping mode) on the WMID to start and stop the data recording, respectively. (Color version available online.)
large number of devices already exist and widely represent the state of the art of inertial monitoring systems,15-18 a custom-made medical platform was developed for this specific framework because of the need to have an open platform for accurately designing and tuning the hardware and software in collaboration with physicians, with a final aim of developing a novel diagnostic tool in sexual medicine. HuMOVE focuses on precise sexual performance medical specifications in terms of derived indexes, intuitiveness, comfort, and confidentiality issues. HuMOVE is intended for SI use by a male human operator and ensures compatibility and minimum physical and psychological interference with the sexual activity to be monitored. The platform derives motion parameters for precise quantization of the physical sexual activity based on the sensed acceleration data. HuMOVE can be exploited for diagnosing sexual pathologies, such as premature or delayed ejaculation, which are the most common male sexual disorders,13,19-23 potentially representing a useful instrument for accurate diagnosis in supporting physicians in the monitoring and long-term UROLOGY 84 (4), 2014
follow-up and pharmaceutical or educational treatment activities.
MATERIALS AND METHODS System Overview An inertial monitoring platform, named HuMOVE, was designed and developed for determining human inertial parameters in sexual performance assessment; it consists of a lowinvasive miniaturized WMID, a magnetic-based trigger (MBT; Fig. 1A), and a high-level data analysis remote station. In the clinical practice of SI, the WMID will be attached to the human back with a patch and the activation and deactivation phases will be comfortably managed by the MBT used as a magneticbased remote control. The WMID is composed of a rectangular-shaped rigid protective package, a custom-made electronic board, and a rechargeable battery; it is 60 mm in length and 36 mm in width and has an overall thickness of 8 mm. The electronic system comprised a miniaturized flexible rectangular-shaped electronic board, which includes a microcontroller, a triaxial digital accelerometer sensor, a memory card 977
for data storage, a magnetic sensor, a magnetic-based switch for the control stages, and light-emitting diodes (LEDs) as the control feedback for the user. Regarding the software implementation, a specific low-level firmware was developed and implemented in the embedded microcontroller to record the data from the inertial sensor in the memory card, to manage the power and activation stages of the device, and to provide correct feedback to the user. The MBT comprised a permanent magnet integrated into a credit cardeshaped case and is used for triggering status changes of the system by simply placing the card close to the WMID. A high-level software data analysis remote control station was developed and implemented on a personal computer for sensor data analysis, parameter extraction, and data organization. An overview of the HuMOVE platform architecture (enclosing the reported modules) is presented in Figure 1B, whereas a detailed description of components and functionalities of the system is provided in the next sections. Electronic System Design. A custom-made miniaturized electronic board was design and developed by integrating off-theshelf components to record and store data from an embedded accelerometer sensor. The board design and specifications were critically investigated because of the dimension issue; the electronic board is 27 mm in length and 36 mm in width and has an overall thickness of about 3.5 mm. The electronic board integrates a microcontroller, that is, the logic unit, selected for its low power consumption, capability, suitable dimensions, and wireless controllability module24 (wireless capability was selected for future improvement toward data download and realtime monitoring). A digital triaxial accelerometer was embedded to record acceleration data from the device with an output data rate of 100 Hz or 400 Hz (ie, 100 or 400 acquired samples per second) and a variable measurement range of 2g or 8g (g is the gravitational acceleration). A frequency of 100 Hz and a measurement range of 2g were selected for suitability with sexual performance analysis. A memory card was integrated in the electronic board for storing X, Y, and Z orthogonal accelerations with a high data rate, avoiding data or time constraint of a wireless communication. A magnetic sensor and a magnetic-based mechanical switch were integrated into the electronic board for managing system transition stages (power and recording management), both controlled by the MBT, whereas a voltage regulator was installed for power control of peripherals. Finally, 2 multicolored LEDs were installed to provide information to the user on the wearable system status. The overall system is powered by an integrated battery. Software Implementation. Software implementation consists of a low-level firmware, integrated in the embedded microcontroller, and a high-level software data analysis remote control station. Regarding the low-level firmware, a code was developed and embedded into the microcontroller; the code is partitioned into 3 main operating blocks with specific functionalities, respectively, named “sleeping mode,” “waiting mode,” and “recording mode” and triggered by the MBT. Once the system is powered, the installed peripherals (ie, microcontroller, accelerometer, magnetic sensor, and memory card) are initialized and the system starts in “sleeping mode.” In this status, all the internal digital circuits powered by the voltage regulator are turned off. “Sleeping mode” was implemented for achieving overall low power consumption, allowing the inertial
978
system to be stored for a long time (ie, at least 42 months storage in this standby condition). Once the MBT is placed near the WMID (ie, both sides indifferently), the magnetic-based switch component enables a logic status change and the system switches to “waiting mode” status (orange LED visual feedback). In this operating mode, the microcontroller and the magnetic sensor are activated and the system can switch to “recording” or “sleeping mode” depending on the magnetic field polarity of the MBT (green or red side on the MBT; Fig. 1A). Once the MBT is drawn near the WMID from the green side, the microcontroller and all the peripherals are activated, the system switches to “recording mode” (green LED visual feedback), and acceleration data are recorded in the memory card (>5 hours total duration). Finally, once the MBT is drawn near the WMID from the red side, the inertial system shifts into “sleeping mode” again (red LED visual feedback), ready for following activations. Working principle and software architecture workflow are represented in Figure 1C and Figure 1D, respectively. Regarding the high-level software, a dedicated routine with a proper user interface was developed for acceleration data analysis and parameter derivation. X, Y, and Z acceleration data are acquired, and data analysis was performed on the modulus of the signal (for recording combined movements independently by the user’s position) after cancellation of the continuous interference by subtracting the mean value of the signal. Data analysis is aimed to extract significant parameters from the raw data. Parameters were properly indicated by the medical expert as the potentially most representative sexual performance indexes: (1) SI time or IELT (until now estimated with a simple stopwatch), (2) an estimation of SI penetrations, as the number of times the signal exceeds a predefined acceleration threshold, and (3) the signal significant frequency (ie, significant number of penetrations per second of the signal). IELT and penetrations estimation are already assessed in the state of the art of sexual diagnosis10-12,14 but hereby evaluable in a low-invasive manner during SI. IELT is automatically extracted by the accelerometer signal, but for a more accurate estimation, the user will be asked to switch the worn system on just before the beginning of the SI and off after the ejaculation event by the MBT. In the sexual analyses framework, the signal significant frequency parameters represent an estimation of the typical cadence of the SI; it is important to correlate this parameter with human physiological and/or pathologic conditions.
System Validation The HuMOVE platform was preliminary validated in terms of effectiveness and reliability through characterization on a controlled model and in a simulated intercourse condition on a human model with the aim to evaluate and tune the system for a forthcoming and realistic translation in the clinical practice. System Validation on a Controlled Model. The aim of the test is to evaluate the accuracy and reliability of acceleration and frequency calculation by the HuMOVE platform with respect to an external high-precision laser displacement sensor benchmark. The WMID was properly fixed on a controlled anthropomorphic manipulator that was programmed for performing 50 iterative movements along an 8-mm linear section (controlled model test 1: A-delay-B-delay oscillations along a straight path [inset in Fig. 2A]). Oscillations were recorded simultaneously by
UROLOGY 84 (4), 2014
Figure 2. (A) Vibration frequency (Hz; oscillations per second) and (B) acceleration (m/s2) estimated by the external distance sensor and the embedded accelerometer simultaneously. Inset in (A) represents the controlled model test bench, whereas the inset in (B) represents the correlation between the acceleration signal measured directly by the accelerometer sensor and by the displacement sensor after double derivation of data. (C) Band-pass filtered module of the acceleration signal for the oscillatory movements at a frequency of 1 Hz and (D) relative enlargement of a single session (first session of [C]). Sexual intercourse penetration numbers are represented with dots in (C) and (D). (Color version available online.)
the accelerometer (integrated into the WMID [100 Hz frequency and measurement range of 2g]) and by the displacement sensor, varying the delay time in each trial (from 0.2 to 1 seconds with 9 incremental steps of 0.1 seconds and setting the maximum manipulator linear velocity and acceleration to 9.5 m/s and 0.95 m/s2, respectively). Mean frequency and acceleration were obtained after appropriately processing the acquired data. The frequency data were computed with a transform algorithm (fast Fourier transform) for both the displacement and acceleration sensor data; acceleration was measured directly by the accelerometer sensor, whereas double derivation was used for the corresponding displacement data. A similar test was performed with the same test bench varying the operating conditions (ie, setting the manipulator linear velocity and the delay time at 9.5 m/s and 0.5 seconds, respectively, and varying acceleration from 0% to 100% of the maximum acceleration with an incremental step of 10% for each trial [controlled model test 2]). Comparative errors between sensors were calculated and reported in the result section as the mean errors standard deviations (SDs) and ranges of error values. System Validation on a Human Model. An experimental validation on a human model was performed for evaluating and assessing the accuracy and reliability of the platform and software processing for parameter calculation in a controlled and real, even if simulated, intercourse operating condition. The device was tested in a simulated SI condition in which 10 UROLOGY 84 (4), 2014
different nonpathologic asserted male users (average age of 29 years, ranging from 26 to 32 years) were asked to perform 25 supervised movements along a linear path on the sagittal plane in an upright position (200-mm-long A-B-A path). Each trial was repeated 3 times by each candidate with a 0.5 Hz, 0.75 Hz, and 1 Hz movement frequency, respectively (human model test 1, 2, and 3, respectively [the frequencies, ie, oscillations per second, beaten by a time meter, were selected in accordance with the medical team to be potentially representative values]). Informed consent according to the Declaration of Helsinki (BMJ. 1991; 302:1194) and the ethical committee of the Scuola Superiore Sant’Anna was obtained before conducting the experiments. The WMID was fastened with a patch on the lumbar area, approximately between the L2 and L5 vertebral bodies; the device was activated and controlled by the user using the MBT (Fig. 1A,C,D). The specific area was selected for the low amount of fat (the position of the device is not related to the body mass index parameter) and its sensitivity to the oscillatory movement. Because of the imposed experimental conditions, an IELT of 50, 33, and 25 seconds; SI penetrations of 25, 25, and 25; and a signal significant frequency of 0.5 Hz, 0.75 Hz, and 1 Hz are expected for the human model test 1, 2, and 3, respectively. The raw data were analyzed deriving the mean values SDs, medians, ranges of values, and mean errors SDs (in comparison with the aforementioned expected outcomes) for the IELT, estimation of SI penetrations, and signal significant frequency. Finally, a questionnaire was administered to the candidates for 979
Figure 3. Bar graphs representing the means standard deviations (SDs) for (A) IELT parameter, (B) SI penetrations, and (C) signal significant frequency for the human model (Hm) test 1, 2, and 3. (D) Detailed numerical results of the Hm test are reported in the table within the figure as means SDs, median, ranges of values, and mean errors SDs with respect to the expected outcomes (within the brackets). d.u., dimensionless unit; IELT, intravaginal ejaculation latency time; SI, sexual intercourse. (Color version available online.)
understanding the acceptance and invasiveness of the overall platform in the simulated SI.
RESULTS System Validation on a Controlled Model A consistent correlation between the accelerometer and displacement sensor data was observed during the controlled model test 1 (as represented by the curves in Fig. 2A,B), demonstrating that the entire developed monitoring platform can provide reliable estimation for the frequency and acceleration of WMID actual oscillations needed for the parameter assessment (mean errors of 0.0102 0.0082 Hz and 0.0389 0.0182 m/s2 and error range values of 0.0008-0.0183 Hz and 0.00020.0389 m/s2, respectively). Similar results were also obtained during the controlled model test 2, and an accurate and reliable correlation between the accelerometer and displacement sensor was once more confirmed (mean errors of 0.0078 0.0051 Hz and 0.0107 0.0066 m/s2 and error range values of 0.00020.0175 Hz and 0.0383-0.0647 m/s2 for frequency and acceleration, respectively). System Validation on a Human Model Results on a human model assessed the effectiveness of the sexual performance evaluation platform and extracted parameters. The software routine figured out data values with a high accuracy for IELT parameters, SI penetrations, and signal significant frequencies for the human model test 1, 2, and 3 in comparison with the expected outcomes imposed to the experiment conditions. Typical acceleration signal representation and SI penetration analysis are depicted in Figure 2C,D, whereas details about the mean values SDs (also represented with the bar graphs), median, ranges of 980
values, and mean errors SDs for the derived parameters are depicted in Figure 3. Obtained outcomes also confirmed the project expectations in terms of usability and comfort, evidenced by the questionnaires that highlighted the low invasiveness and acceptance of the device.
COMMENT Diagnosis of sexual disorders in terms of premature or delayed ejaculation suggests and requires the use of a benchmark for measurement. Today, basic and/or even invasive methodologies, often not compatible with SI, are used for diagnosing sexual disorders (eg, premature ejaculation), such as simple stopwatches (mainly used for assessing IELT parameters), invasive buttoncounter for calculating SI penetrations, and personal evaluation forms about the satisfaction level of the partner.10-14 The major considered issues for the diagnosis of sexual dysfunctions are their quantitative definition and assessment by noninvasive measurement during sexual activity. Along this line, an accelerometer-based custom-made wearable monitoring platform, HuMOVE, was developed with the aim to sense, record, and figure out inertial-based human movement parameters for objectively and accurately assessing human sexual performances with the potential to evaluate sexual dysfunctions, such as premature or delayed ejaculation, in a low-invasive and comfortable fashion during SI.
CONCLUSION An open-platform, adaptable and tunable in terms of hardware, firmware, and data processing software, was developed for flexibility in the implementation of solutions in the sexual medical research field. In this regard, UROLOGY 84 (4), 2014
the development of a custom-made platform guaranteed and will further guarantee tuning and customization through the interaction between the engineering and medical teams, with the final aim to develop a novel suitable biomedical tool for sexual disorder diagnosis. The characterization analysis in the controlled model demonstrated the reliability and accuracy of the development platform in the estimation of the frequency and acceleration of the movement. The system validation on human model in the sexual simulated condition confirmed the effectiveness in the movement assessment and the prospective appropriateness of the system in the sexual medicine scenario as a potential tool for the physicians to help evaluate sexual performance, diagnose disorders, and administer treatment (ie, for premature or delayed ejaculation dysfunctions). The HuMOVE platform also demonstrated to be a comfortable, low-invasive, and easy-to-use measurement system with the potential to be adopted as a standard tool for the sexual medical assessment; it has the benefit of being used during sexual activity with the partner with low psychological effect. Future work will include the performance of experimental investigations in real operating environment on large cohorts of nonpathologic and pathologic patients (also undergoing a pharmaceutical of educational treatment activity) for evaluating the efficacy, suitability, low invasiveness, and acceptance of the system for human sexual performance assessment and dysfunction diagnosis. Moreover, the experimental investigation would also guarantee (1) improvement of the measurement system by finely tuning and/or fixing the most representative parameters (eg, the signal significant frequency inertial-based parameter) for effective analysis investigation, (2) definition of univocal correlations with the pathologic conditions of the sexual activity, and (3) potential introduction of new diagnostic and therapeutic criteria and rules. Finally, thanks to the insights on sexual disorders acquired by the HuMOVE platform, it would be conceivable to develop a therapy system based on a controlled and sensed squeeze action of the urethra duct triggered by acceleration and/or pressure feedback. Acknowledgment. The authors acknowledge the contributions of all the candidates in the testing phase of this work and thank N. Funaro, R. Lazzarini, and R. Di Leonardo for their help in system development.
4.
5.
6. 7. 8.
9.
10.
11.
12.
13.
14. 15.
16. 17.
18.
19. 20.
21.
References 1. Brazzelli M, Saunders DH, Greig CA, Mead GE. Physical fitness training for patients with stroke: an updated review. Stroke. 2012;43: e39-e40. 2. Bricon-Souf N, Newman CR. Context awareness in health care: a review. Int J Med Inform. 2007;76:2-12. 3. Liese AD, Ma X, Maahs DM, Trilk JL. Physical activity, sedentary behaviors, physical fitness, and their relation to health outcomes in
UROLOGY 84 (4), 2014
22. 23. 24.
youth with type 1 and type 2 diabetes: a review of the epidemiologic literature. J Sport Health Sci. 2013;2:21-38. Rauner A, Mess F, Woll A. The relationship between physical activity, physical fitness and overweight in adolescents: a systematic review of studies published in or after 2000. BMC Pediatr. 2013;13: 13-19. Rehn TA, Winett RA, Wisløff U, Rognmo Ø. Increasing physical activity of high intensity to reduce the prevalence of chronic diseases and improve public health. Open Cardiovasc Med J. 2013;7: 1-8. Thompson WR. Worldwide survey of fitness trends for 2013. ACSM’s Health Fitness J. 2012;16:8-17. Bourke A, Lyons G. A threshold-based fall-detection algorithm using a bi-axial gyroscope sensor. Med Eng Phys. 2008;30:84-90. Godfrey A, Conway R, Meagher D, OLaighin G. Direct measurement of human movement by accelerometry. Med Eng Phys. 2008; 30:1364-1386. Curone D, Bertolotti GM, Cristiani A, et al. A real-time and selfcalibrating algorithm based on triaxial accelerometer signals for the detection of human posture and activity. IEEE Trans Inf Technol Biomed. 2010;14:1098-1105. Bar-Or D, Salottolo KM, Orlando A, Winkler JV. A randomized double-blind, placebo-controlled multicenter study to evaluate the efficacy and safety of two doses of the tramadol orally disintegrating tablet for the treatment of premature ejaculation within less than 2 minutes. Eur Urol. 2012;61:736-743. Dinsmore WW, Ralph DJ, Kell P, et al. Evaluation of the Sexual Assessment Monitor, a diagnostic device used to electronically quantify ejaculatory latency time: findings from three studies. BJU Int. 2006;98:613-618. Symonds T, Perelman MA, Althof S, et al. Development and validation of a premature ejaculation diagnostic tool. Eur Urol. 2007;52:565-573. Waldinger M. Recent advances in the classification, neurobiology and treatment of premature ejaculation. Adv Psychosom Med. 2008; 29:50-69. Zargooshi J. Premature ejaculation: bother and intravaginal ejaculatory latency time in Iran. J Sex Med. 2009;6:3478-3489. Aquilano M, Cavallo F, Bonaccorsi M, et al. Ambient assisted living and ageing: preliminary results of RITA project. Conf Proc IEEE Eng Med Biol Soc. 2012;2012:5823-5826. Kahn, Philippe, and Arthur Kinsolving. “Sleep Monitoring System.” U.S. Patent Application 13/545,963. Lynch JP, Loh KJ. A summary review of wireless sensors and sensor networks for structural health monitoring. Shock Vibration Dig. 2006;38:91-130. Tjulkins F, Nguyen A, Andersen M, Imenes K, eds. MEMS accelerometer-based heart monitoring system with myocardial fixation. International Symposium on Biomedical Engineering and Medical Physics, 10-12 October, 2012. Riga, Latvia; 2013:19-22. Gurkan L, Oommen M, Hellstrom WJ. Premature ejaculation: current and future treatments. Asian J Androl. 2008;10:102-109. McMahon CG. Dapoxetine: a new option in the medical management of premature ejaculation. Ther Adv Urol. 2012;4: 233-251. Pastore A, Palleschi G, Leto A, et al. A prospective randomized study to compare pelvic floor rehabilitation and dapoxetine for treatment of lifelong premature ejaculation. Int J Androl. 2012;35: 528-533. Steggall MJ, Pryce A. Premature ejaculation: defining sex in the absence of context. J Men’s Health Gend. 2006;3:25-32. Wyllie MG, Powell JA. The role of local anaesthetics in premature ejaculation. BJU Int. 2012;110:E943-E948. Valdastri P, Menciassi A, Dario P. Transmission power requirements for novel ZigBee implants in the gastrointestinal tract. IEEE Trans Biomed Eng. 2008;55:1705-1710.
981
MATERIALS AND METHODS
RESULTS
CONCLUSION
To investigate an accelerometer-based wearable system, named Human Movement (HuMOVE) platform, designed to enable quantitative and continuous measurement of sexual performance with minimal invasiveness and inconvenience for users. Design, implementation, and development of HuMOVE, a wearable platform equipped with an accelerometer sensor for monitoring inertial parameters for sexual performance assessment and diagnosis, were performed. The system enables quantitative measurement of movement parameters during sexual intercourse, meeting the requirements of wearability, data storage, sampling rate, and interfacing methods, which are fundamental for human sexual intercourse performance analysis. HuMOVE was validated through characterization using a controlled experimental test bench and evaluated in a human model during simulated sexual intercourse conditions. HuMOVE demonstrated to be a robust and quantitative monitoring platform and a reliable candidate for sexual performance evaluation and diagnosis. Characterization analysis on the controlled experimental test bench demonstrated an accurate correlation between the HuMOVE system and data from a reference displacement sensor. Experimental tests in the human model during simulated intercourse conditions confirmed the accuracy of the sexual performance evaluation platform and the effectiveness of the selected and derived parameters. The obtained outcomes also established the project expectations in terms of usability and comfort, evidenced by the questionnaires that highlighted the low invasiveness and acceptance of the device. To the best of our knowledge, HuMOVE platform is the first device for human sexual performance analysis compatible with sexual intercourse; the system has the potential to be a helpful tool for physicians to accurately classify sexual disorders, such as premature or delayed ejaculation. UROLOGY 84: 976e981, 2014. 2014 Elsevier Inc.
O
ver the past few decades, the increased level of awareness within health and physical fitness communities has created an emerging need for smart applications and comfortable wearable devices that sense, classify, and provide information about users’ activities of daily living in a pervasive fashion.1-6 In particular, physical motion, which is a fundamental aspect for categorizing a person’s current health and activity, can be detected and recorded by using wearable accelerometer-based systems.7,8 Their recent impact is mainly because of their small size, low invasiveness, intuitiveness, relatively low cost, and easy integration with existing platforms for sensor networks. They are also able to effectively provide indicative features of human
Financial Disclosure: The authors declare that they have no relevant financial interests. From the The BioRobotics Institute, Scuola Superiore Sant’Anna, Pisa, Italy; the STORM Lab, Department of Mechanical Engineering, Vanderbilt University, Nashville, TN; and the Department of Clinical and Experimental Medicine, University of Pisa, Pisa, Italy Reprint requests: Gastone Ciuti, Ph.D., The BioRobotics Institute, Scuola Superiore Sant’Anna, Viale Rinaldo Piaggio, 34, Pontedera 56025, Pisa, Italy. E-mail: g.ciuti@ sssup.it Submitted: February 11, 2014, accepted (with revisions): June 27, 2014
976
ª 2014 Elsevier Inc. All Rights Reserved
movement corresponding to specific parameters of motion that are important to assess, for instance, static posture characteristics.9 An important field that may benefit from the use of a wearable monitoring inertial device (WMID) is sexual medicine, because of its consistent impact and influence on human life quality and the lack of quantitative and low-invasive monitoring systems for objective parameter analysis during sexual intercourse (SI). Nowadays, basic and/or even invasive methodologies often not compatible with SI are used for diagnosing sexual disorders (eg, premature ejaculation), such as simple stopwatches (mainly used for assessing the intravaginal ejaculation latency time parameter [IELT]), invasive button-counter for calculating SI penetrations, and personal evaluation forms about the satisfaction level of the partner.10-14 Basing on the aforementioned premises, a wearable inertial monitoring platform based on accelerometer sensing, suggested and named Human Movement (HuMOVE) by the author C.B.F., was developed by the authors with the aim to sense, record, and derive inertial movement parameters for evaluating human sexual performance for diagnostic purposes. Although a http://dx.doi.org/10.1016/j.urology.2014.06.040 0090-4295/14
Figure 1. (A) HuMOVE: an accelerometer-based analysis solution for human sexual performance evaluation. (B) System architecture of the overall system composed of a WMID, an MBT, and a remote station. (C) Wearable monitoring inertial device positioned on the human body and working principle and (D) software workflow: once the system is activated and initialized in the “waiting mode” from the “sleeping mode” (phase 1), the user has to approach the MBT with the different sides (phase 2 for recording mode and phase 3 for sleeping mode) on the WMID to start and stop the data recording, respectively. (Color version available online.)
large number of devices already exist and widely represent the state of the art of inertial monitoring systems,15-18 a custom-made medical platform was developed for this specific framework because of the need to have an open platform for accurately designing and tuning the hardware and software in collaboration with physicians, with a final aim of developing a novel diagnostic tool in sexual medicine. HuMOVE focuses on precise sexual performance medical specifications in terms of derived indexes, intuitiveness, comfort, and confidentiality issues. HuMOVE is intended for SI use by a male human operator and ensures compatibility and minimum physical and psychological interference with the sexual activity to be monitored. The platform derives motion parameters for precise quantization of the physical sexual activity based on the sensed acceleration data. HuMOVE can be exploited for diagnosing sexual pathologies, such as premature or delayed ejaculation, which are the most common male sexual disorders,13,19-23 potentially representing a useful instrument for accurate diagnosis in supporting physicians in the monitoring and long-term UROLOGY 84 (4), 2014
follow-up and pharmaceutical or educational treatment activities.
MATERIALS AND METHODS System Overview An inertial monitoring platform, named HuMOVE, was designed and developed for determining human inertial parameters in sexual performance assessment; it consists of a lowinvasive miniaturized WMID, a magnetic-based trigger (MBT; Fig. 1A), and a high-level data analysis remote station. In the clinical practice of SI, the WMID will be attached to the human back with a patch and the activation and deactivation phases will be comfortably managed by the MBT used as a magneticbased remote control. The WMID is composed of a rectangular-shaped rigid protective package, a custom-made electronic board, and a rechargeable battery; it is 60 mm in length and 36 mm in width and has an overall thickness of 8 mm. The electronic system comprised a miniaturized flexible rectangular-shaped electronic board, which includes a microcontroller, a triaxial digital accelerometer sensor, a memory card 977
for data storage, a magnetic sensor, a magnetic-based switch for the control stages, and light-emitting diodes (LEDs) as the control feedback for the user. Regarding the software implementation, a specific low-level firmware was developed and implemented in the embedded microcontroller to record the data from the inertial sensor in the memory card, to manage the power and activation stages of the device, and to provide correct feedback to the user. The MBT comprised a permanent magnet integrated into a credit cardeshaped case and is used for triggering status changes of the system by simply placing the card close to the WMID. A high-level software data analysis remote control station was developed and implemented on a personal computer for sensor data analysis, parameter extraction, and data organization. An overview of the HuMOVE platform architecture (enclosing the reported modules) is presented in Figure 1B, whereas a detailed description of components and functionalities of the system is provided in the next sections. Electronic System Design. A custom-made miniaturized electronic board was design and developed by integrating off-theshelf components to record and store data from an embedded accelerometer sensor. The board design and specifications were critically investigated because of the dimension issue; the electronic board is 27 mm in length and 36 mm in width and has an overall thickness of about 3.5 mm. The electronic board integrates a microcontroller, that is, the logic unit, selected for its low power consumption, capability, suitable dimensions, and wireless controllability module24 (wireless capability was selected for future improvement toward data download and realtime monitoring). A digital triaxial accelerometer was embedded to record acceleration data from the device with an output data rate of 100 Hz or 400 Hz (ie, 100 or 400 acquired samples per second) and a variable measurement range of 2g or 8g (g is the gravitational acceleration). A frequency of 100 Hz and a measurement range of 2g were selected for suitability with sexual performance analysis. A memory card was integrated in the electronic board for storing X, Y, and Z orthogonal accelerations with a high data rate, avoiding data or time constraint of a wireless communication. A magnetic sensor and a magnetic-based mechanical switch were integrated into the electronic board for managing system transition stages (power and recording management), both controlled by the MBT, whereas a voltage regulator was installed for power control of peripherals. Finally, 2 multicolored LEDs were installed to provide information to the user on the wearable system status. The overall system is powered by an integrated battery. Software Implementation. Software implementation consists of a low-level firmware, integrated in the embedded microcontroller, and a high-level software data analysis remote control station. Regarding the low-level firmware, a code was developed and embedded into the microcontroller; the code is partitioned into 3 main operating blocks with specific functionalities, respectively, named “sleeping mode,” “waiting mode,” and “recording mode” and triggered by the MBT. Once the system is powered, the installed peripherals (ie, microcontroller, accelerometer, magnetic sensor, and memory card) are initialized and the system starts in “sleeping mode.” In this status, all the internal digital circuits powered by the voltage regulator are turned off. “Sleeping mode” was implemented for achieving overall low power consumption, allowing the inertial
978
system to be stored for a long time (ie, at least 42 months storage in this standby condition). Once the MBT is placed near the WMID (ie, both sides indifferently), the magnetic-based switch component enables a logic status change and the system switches to “waiting mode” status (orange LED visual feedback). In this operating mode, the microcontroller and the magnetic sensor are activated and the system can switch to “recording” or “sleeping mode” depending on the magnetic field polarity of the MBT (green or red side on the MBT; Fig. 1A). Once the MBT is drawn near the WMID from the green side, the microcontroller and all the peripherals are activated, the system switches to “recording mode” (green LED visual feedback), and acceleration data are recorded in the memory card (>5 hours total duration). Finally, once the MBT is drawn near the WMID from the red side, the inertial system shifts into “sleeping mode” again (red LED visual feedback), ready for following activations. Working principle and software architecture workflow are represented in Figure 1C and Figure 1D, respectively. Regarding the high-level software, a dedicated routine with a proper user interface was developed for acceleration data analysis and parameter derivation. X, Y, and Z acceleration data are acquired, and data analysis was performed on the modulus of the signal (for recording combined movements independently by the user’s position) after cancellation of the continuous interference by subtracting the mean value of the signal. Data analysis is aimed to extract significant parameters from the raw data. Parameters were properly indicated by the medical expert as the potentially most representative sexual performance indexes: (1) SI time or IELT (until now estimated with a simple stopwatch), (2) an estimation of SI penetrations, as the number of times the signal exceeds a predefined acceleration threshold, and (3) the signal significant frequency (ie, significant number of penetrations per second of the signal). IELT and penetrations estimation are already assessed in the state of the art of sexual diagnosis10-12,14 but hereby evaluable in a low-invasive manner during SI. IELT is automatically extracted by the accelerometer signal, but for a more accurate estimation, the user will be asked to switch the worn system on just before the beginning of the SI and off after the ejaculation event by the MBT. In the sexual analyses framework, the signal significant frequency parameters represent an estimation of the typical cadence of the SI; it is important to correlate this parameter with human physiological and/or pathologic conditions.
System Validation The HuMOVE platform was preliminary validated in terms of effectiveness and reliability through characterization on a controlled model and in a simulated intercourse condition on a human model with the aim to evaluate and tune the system for a forthcoming and realistic translation in the clinical practice. System Validation on a Controlled Model. The aim of the test is to evaluate the accuracy and reliability of acceleration and frequency calculation by the HuMOVE platform with respect to an external high-precision laser displacement sensor benchmark. The WMID was properly fixed on a controlled anthropomorphic manipulator that was programmed for performing 50 iterative movements along an 8-mm linear section (controlled model test 1: A-delay-B-delay oscillations along a straight path [inset in Fig. 2A]). Oscillations were recorded simultaneously by
UROLOGY 84 (4), 2014
Figure 2. (A) Vibration frequency (Hz; oscillations per second) and (B) acceleration (m/s2) estimated by the external distance sensor and the embedded accelerometer simultaneously. Inset in (A) represents the controlled model test bench, whereas the inset in (B) represents the correlation between the acceleration signal measured directly by the accelerometer sensor and by the displacement sensor after double derivation of data. (C) Band-pass filtered module of the acceleration signal for the oscillatory movements at a frequency of 1 Hz and (D) relative enlargement of a single session (first session of [C]). Sexual intercourse penetration numbers are represented with dots in (C) and (D). (Color version available online.)
the accelerometer (integrated into the WMID [100 Hz frequency and measurement range of 2g]) and by the displacement sensor, varying the delay time in each trial (from 0.2 to 1 seconds with 9 incremental steps of 0.1 seconds and setting the maximum manipulator linear velocity and acceleration to 9.5 m/s and 0.95 m/s2, respectively). Mean frequency and acceleration were obtained after appropriately processing the acquired data. The frequency data were computed with a transform algorithm (fast Fourier transform) for both the displacement and acceleration sensor data; acceleration was measured directly by the accelerometer sensor, whereas double derivation was used for the corresponding displacement data. A similar test was performed with the same test bench varying the operating conditions (ie, setting the manipulator linear velocity and the delay time at 9.5 m/s and 0.5 seconds, respectively, and varying acceleration from 0% to 100% of the maximum acceleration with an incremental step of 10% for each trial [controlled model test 2]). Comparative errors between sensors were calculated and reported in the result section as the mean errors standard deviations (SDs) and ranges of error values. System Validation on a Human Model. An experimental validation on a human model was performed for evaluating and assessing the accuracy and reliability of the platform and software processing for parameter calculation in a controlled and real, even if simulated, intercourse operating condition. The device was tested in a simulated SI condition in which 10 UROLOGY 84 (4), 2014
different nonpathologic asserted male users (average age of 29 years, ranging from 26 to 32 years) were asked to perform 25 supervised movements along a linear path on the sagittal plane in an upright position (200-mm-long A-B-A path). Each trial was repeated 3 times by each candidate with a 0.5 Hz, 0.75 Hz, and 1 Hz movement frequency, respectively (human model test 1, 2, and 3, respectively [the frequencies, ie, oscillations per second, beaten by a time meter, were selected in accordance with the medical team to be potentially representative values]). Informed consent according to the Declaration of Helsinki (BMJ. 1991; 302:1194) and the ethical committee of the Scuola Superiore Sant’Anna was obtained before conducting the experiments. The WMID was fastened with a patch on the lumbar area, approximately between the L2 and L5 vertebral bodies; the device was activated and controlled by the user using the MBT (Fig. 1A,C,D). The specific area was selected for the low amount of fat (the position of the device is not related to the body mass index parameter) and its sensitivity to the oscillatory movement. Because of the imposed experimental conditions, an IELT of 50, 33, and 25 seconds; SI penetrations of 25, 25, and 25; and a signal significant frequency of 0.5 Hz, 0.75 Hz, and 1 Hz are expected for the human model test 1, 2, and 3, respectively. The raw data were analyzed deriving the mean values SDs, medians, ranges of values, and mean errors SDs (in comparison with the aforementioned expected outcomes) for the IELT, estimation of SI penetrations, and signal significant frequency. Finally, a questionnaire was administered to the candidates for 979
Figure 3. Bar graphs representing the means standard deviations (SDs) for (A) IELT parameter, (B) SI penetrations, and (C) signal significant frequency for the human model (Hm) test 1, 2, and 3. (D) Detailed numerical results of the Hm test are reported in the table within the figure as means SDs, median, ranges of values, and mean errors SDs with respect to the expected outcomes (within the brackets). d.u., dimensionless unit; IELT, intravaginal ejaculation latency time; SI, sexual intercourse. (Color version available online.)
understanding the acceptance and invasiveness of the overall platform in the simulated SI.
RESULTS System Validation on a Controlled Model A consistent correlation between the accelerometer and displacement sensor data was observed during the controlled model test 1 (as represented by the curves in Fig. 2A,B), demonstrating that the entire developed monitoring platform can provide reliable estimation for the frequency and acceleration of WMID actual oscillations needed for the parameter assessment (mean errors of 0.0102 0.0082 Hz and 0.0389 0.0182 m/s2 and error range values of 0.0008-0.0183 Hz and 0.00020.0389 m/s2, respectively). Similar results were also obtained during the controlled model test 2, and an accurate and reliable correlation between the accelerometer and displacement sensor was once more confirmed (mean errors of 0.0078 0.0051 Hz and 0.0107 0.0066 m/s2 and error range values of 0.00020.0175 Hz and 0.0383-0.0647 m/s2 for frequency and acceleration, respectively). System Validation on a Human Model Results on a human model assessed the effectiveness of the sexual performance evaluation platform and extracted parameters. The software routine figured out data values with a high accuracy for IELT parameters, SI penetrations, and signal significant frequencies for the human model test 1, 2, and 3 in comparison with the expected outcomes imposed to the experiment conditions. Typical acceleration signal representation and SI penetration analysis are depicted in Figure 2C,D, whereas details about the mean values SDs (also represented with the bar graphs), median, ranges of 980
values, and mean errors SDs for the derived parameters are depicted in Figure 3. Obtained outcomes also confirmed the project expectations in terms of usability and comfort, evidenced by the questionnaires that highlighted the low invasiveness and acceptance of the device.
COMMENT Diagnosis of sexual disorders in terms of premature or delayed ejaculation suggests and requires the use of a benchmark for measurement. Today, basic and/or even invasive methodologies, often not compatible with SI, are used for diagnosing sexual disorders (eg, premature ejaculation), such as simple stopwatches (mainly used for assessing IELT parameters), invasive buttoncounter for calculating SI penetrations, and personal evaluation forms about the satisfaction level of the partner.10-14 The major considered issues for the diagnosis of sexual dysfunctions are their quantitative definition and assessment by noninvasive measurement during sexual activity. Along this line, an accelerometer-based custom-made wearable monitoring platform, HuMOVE, was developed with the aim to sense, record, and figure out inertial-based human movement parameters for objectively and accurately assessing human sexual performances with the potential to evaluate sexual dysfunctions, such as premature or delayed ejaculation, in a low-invasive and comfortable fashion during SI.
CONCLUSION An open-platform, adaptable and tunable in terms of hardware, firmware, and data processing software, was developed for flexibility in the implementation of solutions in the sexual medical research field. In this regard, UROLOGY 84 (4), 2014
the development of a custom-made platform guaranteed and will further guarantee tuning and customization through the interaction between the engineering and medical teams, with the final aim to develop a novel suitable biomedical tool for sexual disorder diagnosis. The characterization analysis in the controlled model demonstrated the reliability and accuracy of the development platform in the estimation of the frequency and acceleration of the movement. The system validation on human model in the sexual simulated condition confirmed the effectiveness in the movement assessment and the prospective appropriateness of the system in the sexual medicine scenario as a potential tool for the physicians to help evaluate sexual performance, diagnose disorders, and administer treatment (ie, for premature or delayed ejaculation dysfunctions). The HuMOVE platform also demonstrated to be a comfortable, low-invasive, and easy-to-use measurement system with the potential to be adopted as a standard tool for the sexual medical assessment; it has the benefit of being used during sexual activity with the partner with low psychological effect. Future work will include the performance of experimental investigations in real operating environment on large cohorts of nonpathologic and pathologic patients (also undergoing a pharmaceutical of educational treatment activity) for evaluating the efficacy, suitability, low invasiveness, and acceptance of the system for human sexual performance assessment and dysfunction diagnosis. Moreover, the experimental investigation would also guarantee (1) improvement of the measurement system by finely tuning and/or fixing the most representative parameters (eg, the signal significant frequency inertial-based parameter) for effective analysis investigation, (2) definition of univocal correlations with the pathologic conditions of the sexual activity, and (3) potential introduction of new diagnostic and therapeutic criteria and rules. Finally, thanks to the insights on sexual disorders acquired by the HuMOVE platform, it would be conceivable to develop a therapy system based on a controlled and sensed squeeze action of the urethra duct triggered by acceleration and/or pressure feedback. Acknowledgment. The authors acknowledge the contributions of all the candidates in the testing phase of this work and thank N. Funaro, R. Lazzarini, and R. Di Leonardo for their help in system development.
4.
5.
6. 7. 8.
9.
10.
11.
12.
13.
14. 15.
16. 17.
18.
19. 20.
21.
References 1. Brazzelli M, Saunders DH, Greig CA, Mead GE. Physical fitness training for patients with stroke: an updated review. Stroke. 2012;43: e39-e40. 2. Bricon-Souf N, Newman CR. Context awareness in health care: a review. Int J Med Inform. 2007;76:2-12. 3. Liese AD, Ma X, Maahs DM, Trilk JL. Physical activity, sedentary behaviors, physical fitness, and their relation to health outcomes in
UROLOGY 84 (4), 2014
22. 23. 24.
youth with type 1 and type 2 diabetes: a review of the epidemiologic literature. J Sport Health Sci. 2013;2:21-38. Rauner A, Mess F, Woll A. The relationship between physical activity, physical fitness and overweight in adolescents: a systematic review of studies published in or after 2000. BMC Pediatr. 2013;13: 13-19. Rehn TA, Winett RA, Wisløff U, Rognmo Ø. Increasing physical activity of high intensity to reduce the prevalence of chronic diseases and improve public health. Open Cardiovasc Med J. 2013;7: 1-8. Thompson WR. Worldwide survey of fitness trends for 2013. ACSM’s Health Fitness J. 2012;16:8-17. Bourke A, Lyons G. A threshold-based fall-detection algorithm using a bi-axial gyroscope sensor. Med Eng Phys. 2008;30:84-90. Godfrey A, Conway R, Meagher D, OLaighin G. Direct measurement of human movement by accelerometry. Med Eng Phys. 2008; 30:1364-1386. Curone D, Bertolotti GM, Cristiani A, et al. A real-time and selfcalibrating algorithm based on triaxial accelerometer signals for the detection of human posture and activity. IEEE Trans Inf Technol Biomed. 2010;14:1098-1105. Bar-Or D, Salottolo KM, Orlando A, Winkler JV. A randomized double-blind, placebo-controlled multicenter study to evaluate the efficacy and safety of two doses of the tramadol orally disintegrating tablet for the treatment of premature ejaculation within less than 2 minutes. Eur Urol. 2012;61:736-743. Dinsmore WW, Ralph DJ, Kell P, et al. Evaluation of the Sexual Assessment Monitor, a diagnostic device used to electronically quantify ejaculatory latency time: findings from three studies. BJU Int. 2006;98:613-618. Symonds T, Perelman MA, Althof S, et al. Development and validation of a premature ejaculation diagnostic tool. Eur Urol. 2007;52:565-573. Waldinger M. Recent advances in the classification, neurobiology and treatment of premature ejaculation. Adv Psychosom Med. 2008; 29:50-69. Zargooshi J. Premature ejaculation: bother and intravaginal ejaculatory latency time in Iran. J Sex Med. 2009;6:3478-3489. Aquilano M, Cavallo F, Bonaccorsi M, et al. Ambient assisted living and ageing: preliminary results of RITA project. Conf Proc IEEE Eng Med Biol Soc. 2012;2012:5823-5826. Kahn, Philippe, and Arthur Kinsolving. “Sleep Monitoring System.” U.S. Patent Application 13/545,963. Lynch JP, Loh KJ. A summary review of wireless sensors and sensor networks for structural health monitoring. Shock Vibration Dig. 2006;38:91-130. Tjulkins F, Nguyen A, Andersen M, Imenes K, eds. MEMS accelerometer-based heart monitoring system with myocardial fixation. International Symposium on Biomedical Engineering and Medical Physics, 10-12 October, 2012. Riga, Latvia; 2013:19-22. Gurkan L, Oommen M, Hellstrom WJ. Premature ejaculation: current and future treatments. Asian J Androl. 2008;10:102-109. McMahon CG. Dapoxetine: a new option in the medical management of premature ejaculation. Ther Adv Urol. 2012;4: 233-251. Pastore A, Palleschi G, Leto A, et al. A prospective randomized study to compare pelvic floor rehabilitation and dapoxetine for treatment of lifelong premature ejaculation. Int J Androl. 2012;35: 528-533. Steggall MJ, Pryce A. Premature ejaculation: defining sex in the absence of context. J Men’s Health Gend. 2006;3:25-32. Wyllie MG, Powell JA. The role of local anaesthetics in premature ejaculation. BJU Int. 2012;110:E943-E948. Valdastri P, Menciassi A, Dario P. Transmission power requirements for novel ZigBee implants in the gastrointestinal tract. IEEE Trans Biomed Eng. 2008;55:1705-1710.
981
