Permeable electrodes for electrodermal activity

Abstract

An electrodermal activity (EDA) sensor has one or more pairs of permeable electrodes, configured to be located on skin of a user, a voltage source operable to apply a voltage across the electrodes, and a port electrically coupled to the one or more pairs of electrodes to transmit readings from the one or more pairs of permeable electrodes.

Description

PERMEABLE ELECTRODES FOR ELECTRODERMAL ACTIVITYTECHNICAL FIELD

[0001] This disclosure relates to sensors for detection of electrodermal activity, more particularly to sensors using permeable electrodes.BACKGROUND

[0002] Electrodermal activity (EDA) has long been used for mental activity monitoring by measuring skin conductance (SkinG) only at specific locations with high sweat gland density. However, EDA is not considered useful for physical activity because of the shifted baseline due to accumulation of sweat at the skin / electrode interface.

[0003] The autonomic nervous system regulates involuntary physiological processes for homeostasis in response to physical and mental activities. Thus, wearable biosensors for physiological signals such as heart rate, blood pressure, body temperature, or sweat secretion rate can be used for monitoring a person’s physical or mental state. However, psychological activities induce relatively small physiological changes and are difficult to detect compared to physical activities5. Nevertheless, electrodermal activity (EDA) has long been used for psychological stress or excitement by measuring skin conductance (SkinG) at specific locations with high sweat gland density, such as the fingers, soles of the feet, and forehead. During mental activities, SkinG consists of a long-lasting, slowly drifting baseline (tonic, SkinG level) overlaid by short-term, subtle pulses (phasic, SkinG response).

[0004] Because the phasic component represents signaling by the sympathetic nerve system — leading to the filling of and release of the sweat gland — many clinicians have used it to analyze mental activity. Recently, some researchers tried unsuccessfully to use EDA for physical activity monitoring due to the lack of recovery back to the baseline. Once sweat is trapped between hydrated skin and the electrode, the tonic SkinG level does not return to the baseline due to the inevitable accumulation of phasic components on the stratum comeum over long period of time. This challenging recovery issue limits the use of SkinG sensors for continuously monitoring physical activities with high sweat rates and long duration measurements at a low sweat rate.BRIEF DESCRIPTION OF THE DRAWINGS

[0005] FIGa. 1A-1E show images and schematics of an embodiment of a multimodal sweat sensor for monitoring mental and physical activities.

[0006] FIGs. 2A-2E show various embodiments for permeable electrodes for skin conductance (SkinG).

[0007] FIGs. 3A-3D show schematics of various designs of micro-lace electrodes.

[0008] FIGs. 4A-4B show photographs of effects of electrode permeability on user’s fingertips.

[0009] FIGs. 5A-5C show graphs of SkinG recovery responses using a standard plate electrode.

[0010] FIGs. 6A-6B show graphs of qualitative SkinG analysis of mental and physical activities on different body sites.

[0011] FIGs. 7A-7C show graphs of results for an on-body real-time subject study during stationary cycling for quantitative analysis.

[0012] FIG. 8 shows a graph of results from monitoring physical and mental activities with long-term use of a SkinG sensor in accordance with the embodiments.

[0013] FIG. 9 shows an embodiment of an integrated SkinG multiplexed sensing system.

[0014] FIGs. 10A-10D show an embodiment of a SkinG printed circuit board for multiplexed wireless sensing.

[0015] FIG. 11 shows an embodiment of a user interface for a mobile application.DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] The embodiments herein use electrodermal activity (EDA) as a proxy for sweat secretion rate by using water-permeable electrodes, such as p-lace, spiral, or conductive fabric, electrodes. Incorporating micro-lace (p-lace) electrodes and a microfluidic channel, the multimodal sweat sensor could measure skin hydration level and locally secreted sweat rate simultaneously as shown in FIG. 1 A. Smaller than the size of a sweat pore, the microsized pattern on the lace-like electrodes allows SkinG values to recover from increased levels from hydration. Also, multi-site measurements of SkinG enabled continuous monitoring and selective decoupling of physical and mental activities using the natural property that differentskin sites on the body have different sweat gland densities and recruitment. While the finger can respond to physical and mental activities, some sites like the wrist, forearm, and upper arm showed selective response only to physical activity. This SkinG sensing platform is a powerful tool enabling clinical studies with more accurate and reliable EDA analysis, or even in this new platform as a reusable proxy for sweat rate sensors.

[0017] As illustrated in FIG. 1 A, a multimodal sweat sensor consisting of p-lace electrodes 10, 20, and a microfluidic channel allows the continuous and simultaneous measurement of SkinG and sweat rate. While the microfluidic sweat sensor measures the progressing volume of locally secreted sweat, the SkinG system continuously measures with two paired electrodes, 10 and 20 lie on a sweat pore 14 on the skin. As paired electrodes for SkinG sensing, titanium (Ti) and gold (Au) were deposited sequentially on top of a thin substrate of polyethylene terephthalate (PET). Further, the skin-contacting areas of the electrodes were laser-patterned for water-permeability. As in FIG. IB, the SkinG sensing p-lace electrodes of the embodiments can be easily combined with a microfluidic sensor 30 by laminating the pair of electrodes onto the microfluidic. Thin and flexible p-lace electrodes showed conformal contact with the bumpy tendons of the wrist FIG. 1C.

[0018] To relate SkinG to the skin hydration level, one can assume the electrical model of the skin that is illustrated in FIG. ID. In the skin, a sweat gland in the dermis layer continues outward through the epidermis layer as a channel called the dermal duct, ending in a sweat pore 14 on the surface. By applying a voltage across two electrodes directly connected to the skin, one can measure the total electrical resistance (1 / conductance, R) across the regions of electrode contacts 10 and 20, stratum corneum (Asc) 12, epidermis, and dermis, involving sweat glands 18, interstitial fluid, and vascularization. Because the internal Abody across conductive pathways of interstitial fluid in a sweat duct 16 and vascularization in the dermis and deeper inside the body is only hundreds of ohms. These are a few orders smaller than mega ohm level Asc. One can assume that SkinG is approximately 1 / (2RS / ). Because SkinG increases from the dehydrated state to the fully hydrated state of the stratum corneum, one can correlate the total amount of locally secreted sweat to the amount of sweat absorbed on the stratum corneum during the partially hydrated stage. Using this skin hydration level modeling, the embodiments use tonic SkinG of EDA as a proxy for local sweat rate sensors.

[0019] The SkinG sensor of the embodiments can selectively monitor mental and physical activity because different skin sites show different SkinG responses as shown in FIG. IE. Asin typical EDA analysis, one can obtain phasic responses of SkinG from mental stimulus on skin sites with high sweat gland density, such as the distal fingertip phalanges, palms, soles of the feet, and forehead. On the other hand, no SkinG changes are observed for mental stimuli on sites with lower sweat gland densities. When physical activity is conducted, all sites may show increments in SkinG because physical activities can activate sweat glands for thermoregulation, inducing high sweating. However, while sites with the lower gland densities only show a cumulative tonic response, the high-density sites show both tonic and phasic responses with similar amplitude. One can discriminate between mental and physical activity by identifying the typical form of the signals from multi-site measurements.

[0020] As electrodes for SkinG sensing, the embodiments introduce various kinds of water- permeable and conductive electrodes FIG. 2A shows a schematic of non-permeable and permeable electrodes, 40 and 42, respectively. Permeable electrodes, which may include a micro-lace (p-lace) electrode 44, spiral metal wire 46, and carbon fiber 48, or other conductive fabric, electrodes as shown in FIG. 2C. FIGs. 2D and 2E shows graphs comparing the results of micro-lace permeable electrodes versus normal plate electrodes. Each permeable electrode offers much higher water permeability than a non-permeable polymer sheet or metal plate (with 0% permeability). Water permeability was measured for each electrode with three different designs, finding that p-lace and spiral wire present nearly complete water permeability and negligible water absorption, with 99.4% and 98.4% of water flowing through the electrode. Although carbon fiber fabric showed moderate permeability of 89.6% and slightly high water absorption of 11.4%, these fabric-type electrodes may have a specific use case because of their high wearability.

[0021] For direct and conformal contact to the skin, p-lace patterns were optimized further using the computer-assisted laser cutting as in FIG. 3 A shows a schematic pattern of p-lace electrodes, along with pictures of p-lace electrodes with three different lace widths, 200 pm 50, 100 pm 52, and 50 pm 54 (Scale bar, 1 mm) B-D, comparison among three p-lace patterns of water permeability in FIG. 3B, electrical resistance of 12 mm x 55 mm electrodes in FIG. 3C, and relative change of the resistance (AR R ) under stretching in FIG. 3D. As used here, the term “micro-lace” or “p-lace” means a structure consisting of a repeated patten of a geometric shape, such as, but not limited to, hexagons, squares, circles, octagons, triangles, etc. The inventors designed the pattern of close-packed hexagons of spacing 1.27 mm with three different lace widths of approximately 50, 100, and 200 pm as shown in FIG.3 A. The different laces showed permeability of 99.4, 99.0, and 98.2%, respectively, shown in FIG. 3B. To evaluate the electro-mechanical properties of the p-lace electrodes of the embodiments, the inventors compared A of the same patterns in 12 mm x 55 mm rectangular electrodes before and under stretching, shown in FIGs. 3C and 3D. The p-lace electrodes of the embodiments released the strain energy structurally and showed much less relative R change until 8% strain, while solid film electrodes showed over ten times relative R change with plastic deformation.

[0022] As schematically illustrated in FIG. 2B, a non-permeable sheet electrode can trap the sweat and induce swelling of the skin underneath the electrode. In contrast, permeable p-lace electrodes enable the permeation of the sweat through the holes of electrodes and the evaporation of sweat into the air. After 6 hours of use on the fingertip, the non-permeable sheet electrode left a blanching mark on the swollen skin, indicating the sweat was trapped underneath the electrode. In contrast, the p-lace electrode did not leave any mark on the skin after the same period. The inventors also wore plate metal 55, shown in FIG. 4A, and spiral metal wire 46, shown in FIG. 4B, for 6 hours and noticed a blanching mark by the non- permeable plate electrode.

[0023] To verify the electrically beneficial effect of the electrode’s permeability, the inventors measured SkinG on both the finger at rest and the upper arm during and after exercise. In FIG. 2D, one can see much smaller changes in tonic response with the p-lace electrodes, while sheet electrodes showed increasing tonic response although the subject was at rest. In another test, the inventors also tried 15 minutes of riding a stationary bike to measure the recovery response of SkinG on the upper arm. From non-permeable electrodes, such as the one shown in FIG. 5 A, one would expect early saturation of the value during physical activity and less recovery after physical activity. FIG. 5B shows SkinG variations on the fingertip at rest, and FIG. 5C shows SkinG recovery responses on the upper arm after 15 minutes of riding a stationary bike.

[0024] The inventors selected four sites to test various body sites (finger, wrist, forearm, and upper arm). Additionally, they simultaneously attached a commercial activity tracker (such as Apple Watch or Fitbit) and a body temperature sensor (CORE) to the skin for comparison. The self-scored sweat rate was qualitatively assessed on a scale from 0 to 3. A sweat score of 0 corresponds to at-rest sweat rates, while a score of 3 corresponds to heavy sweating, similar to when the subject undergoes intense exercise.

[0025] As shown in FIG. 6A, SkinG on the finger can be used for mental activity monitoring, while the wrist can be used for physical activity monitoring. Stable and peaceful mental states show sparse spikes on the finger SkinG, while nervous and irritated states from mental stimulus can fire frequent repetitive signals. On the wrist, SkinG predominantly shows the tonic response to the amount of sweat from thermoregulation, as in during exercise. Using a microfluidic sweat rate sensor in the multimodal sweat sensor on the wrist, it is possible to qualitatively correlate the SkinG to the sweat rate.

[0026] To apply the mental stimulus in FIG. 6B, the subject conducted an IQ test for 25 minutes and subsequently closed their eyes for 10 minutes, wearing four sensors on the different sites: finger, wrist, forearm, and upper arm. This resulted in varying SkinG signals with firing spikes only from the finger. After mental activities, the inventors processed SkinG with high pass filtering to separate the phasic and tonic responses, as typically done in EDA analysis. In the phasic component of the signal, one can see a difference between the frequent spikes caused by a stressful IQ test and the absence of spikes caused by peaceful eyes closed.

[0027] The SkinG sensor can be comfortably anchored on the skin with a self-adherent sports band on various body parts. Under a two-phase cycling activity with controlled pedaling resistance and fixed speed in FIG. 6B, the SkinG from wrist, forearm, and upper arm showed similar increasing trends as expected, followed by fast recovery to the baseline, which is critical for long-term measurement and reusability of the sensor. Sweat rates appeared nearconstant during exercise periods, while SkinG showed an increasing slope, indicating the hydration level of the skin related to the accumulation of sweat on the skin. On the finger, there were also tonic increases during exercise. However, the amplitude of the tonic increases was similar to that of the phasic spike responses7, meaning that finger SkinG is reflective of both mental and physical activities. Therefore, the sensor can distinguish mental and physical activities using multi-site measurements.

[0028] To quantitatively correlate SkinG and skin hydration levels, the inventors conducted a subject study measuring both SkinG and sweat secretion volume in localized areas continuously over time. As in FIG. 7 A, 18 subjects wore SkinG sensors on the forearm, and at the same time, they also put an established commercial sweat collector on the other forearm to measure local sweat volume. To calibrate for subject-specific differences in skin properties such as skin thickness or ease of sweat absorption, relative changes of SkinG (ASkinG / SkinGo) were calculated by dividing the SkinG variation (ASkinG) by initial baseline SkinG (SkinGo). The subjects exercised on the stationary bike at a constant speed for15 minutes. Riding speed on the stationary bike depended on the subject but was constant throughout the time period.

[0029] For SkinG sensing, the subjects wore a pair of spiral metal wire electrodes only on the right forearm. To quantize the local sweat amount, subjects wore the commercial microfluidic sweat collector (Macroduct) on the left forearm. Subjects were guided to ride the stationary bike for 15 minutes at a comfortable speed. Tests on the same subjects were conducted on different days, and only those who felt comfortable increased their speed. After 15 minutes of exercise, the subject rested for more than 5 minutes. SkinG signals from the finger were decomposed to the phasic SkinG response and tonic SkinG level using the Python package NeuroKit226.

[0030] As in FIG. 7B, the inventors found a strong positive correlation, from 12participants, between relative change of SkinG and local sweat volume (slope: 8.45, R2:0.83). This positive relation indicates that SkinG values are proportional to the amount of sweat secreted and absorbed into the stratum corneum. Although there are still differences among the subjects, it was possible to see higher SkinG values for the same subject in multiple trials with increasing speed, as in FIG. 7C. Volumetric sweat measurements from the commercial sweat collector also increased. Sweat volume data could not be obtained at the slower speeds of 18 km / h and 19.5 km / h due to the dead volume of the commercial sweat collector.)

[0031] Innovation in long-term, continuous tracking of human activities is needed for daily life or athletic studies because it has so far been limited only to a few physiological parameters like heart rate, temperature, and blood pressure, which offer limited information on mental activity. In FIG. 8, the subject wore two SkinG sensors on the finger and the forearm for over 11 hours for continuous and selective monitoring of mental and physical activity during a typical daily routine, including walking, sitting, meeting, eating, and talking. Also, an activity tracker for steps and heart rate, skin and core temperature sensor, and selfassessed sweating scores were used simultaneously.

[0032] Step count and sweating score from walking matched well with increasing forearm SkinG, heart rate, and core temperature. The decreasing skin temperature was due to the lower ambient temperature outside. However, in the finger SkinG, it was not possible to distinguish physical activity from other daily stimuli because finger SkinG had no positive correlation to the sweating score. Instead, finger SkinG exhibited huge conductance changes in response to excitement and psychological stimuli during mental activities like meeting andtalking. In the processed finger SkinG for phasic EDA, many spikes corresponded to the mental stimuli, while the other physiological signals like heart rate or temperature could not identify any differences in mental stimuli. It is believed that SkinG sensing enhances current wearable sensing platforms and has the potential to facilitate additional fundamental physiological studies.

[0033] In one embodiment, the fabrication process involved winding STS wire and copper among the permeable electrodes, the STS wire and copper (Cu) wire were wound to form a spiral pattern for the spiral metal wire electrodes. To make the Cu chemically stable, the inventors electroplated Au on the surface of the Cu wire. The p-lace electrodes were laser patterned using a Trotec Speedy 400 laser running at 5-15% power, 0.3-0.5% speed, and 60 kHz on a 10 pm PET with Ti / Au thin film of 50 / 100 nm, respectively. Carbon fiber fabric was purchased from Techinical Fibre Products Ltd. SkinG sensors were anchored on the skin with a self-adherent sports band (Band-Aid, Johnson and Johnson) for the stable contact. For integrated multimodal sweat sensors, p-lace electrodes were laminated on the inlet side of the microfluidic channel. A detailed description of the fabrication process of microfluidic channels was included in a previous study25.

[0034] The inventors conducted the permeability test by pouring water onto the sample suspended over a container and measuring the mass changes. After removing the sample, they determined the amount of water that passed through. Dividing this by the total amount of poured water, they calculated the permeability without considering fluid pressure.

[0035] SkinG at up to four skin sites was measured continuously via a custom PCB implementing the standard exosomatic direct current (DC) voltage method with wireless readout to a mobile application, as shown in FIGs. 9 and 10 and documented in Table 1, below.

[0036] FIG. 9 shows an embodiment of a printed circuit board 60 shown with an attached rechargeable battery module 64 and 8-conductor connecting cable 62 for up to four pairs of sensing electrodes, one pair of which is shown in the image, electrodes 10 and 20.

[0037] FIG. 10 shows an embodiment of a design of a SkinG printed circuit board for multiplexed wireless sensing. FIG. 10A shows a circuit schematic of conductance measurement channel having a differential amplifier 70 for input and a low pass filter 72, in this embodiment an RC filter. The bottom of FIG. 10A shows a multiplexer 74 used to multiplex signals from 4 pairs of electrodes. FIG. 10B shows calibration of conductancemeasurement from microcontroller ADC (analog-to-digital converter) code readout. FIG.IOC shows power delivery to analog and digital circuitry. The battery 80 provides power to the voltage regulator, which then provides power to the analog channel 84, and the digital channel. In this embodiment, the digital channel comprises a wireless module microcontroller 86 for a BLE (BlueTooth Low Energy), such as may connect to a smart watch or fitness tracker. FIG. 10D shows a diagram of timing for measurement multiplexing, oversampling, and wireless data transmission.

[0038] Table 1 | SkinG board measurement characteristics

[0039] In one embodiment, the input amplifier 70 in FIG. 10A comprises a trans-impedance amplifier (TIA) configured with an LTC2063 operational amplifier, applying a 0.8-1 V voltage between on-skin electrodes, and including a current-limiting resistor preventing the applied current from exceeding 100 pA; note that the limit of human perception is typically taken to be 1 mA.

[0040] Table 2 | SkinG board safety

[0041] Each pair of electrodes is measured at 1Hz with a 250 ms interval between sites, as60 shown in the timing diagram of FIG. 10D . The resolution range is 0-32.7 pS; typical skin conductance falls between 1-20 pS, shown in Table 3.

[0042] Table 3 | SkinG board measurement range

[0043] Readings are transmitted to a mobile phone via Bluetooth Low Energy for data monitoring and export. As in FIG. 11, a phone application was developed for wireless data transfer and extraction. FIG. 11 shows an example of a user interface 90 for an application that can track the electrode outputs.

[0044] The embodiments here show a multimodal sweat sensing platform to locally measure sweat loss with microfluidics and accurate skin hydration levels with SkinG sensors exploiting EDA technology. By implementing water-permeable p-lace electrodes, the sensor can remove the drift in tonic response from sweat accumulation under the device and monitor high sweat rate physical activities. In a subject study, the inventors verified the validity of EDA as a proxy for sweat rate. Also, based on the finding that response to physical and mental stimuli depends on the skin site, the inventors could continuously monitor and selectively decouple physical and mental activities with multi-site SkinG measurements. The readings can be transmitted from the sensors through a wireless port to a receiver such as a mobile phone or control station or transmitted through a wired port to which the sensors are attached.

[0045] All features disclosed in the specification, including the claims, abstract, and drawings, and all the steps in any method or process disclosed, may be combined in any combination, except combinations where at least some of such features and / or steps aremutually exclusive. Each feature disclosed in the specification, including the claims, abstract, and drawings, can be replaced by alternative features serving the same, equivalent, or similar purpose, unless expressly stated otherwise.

[0046] Additionally, this written description makes reference to particular features. It is to be understood that the disclosure in this specification includes all possible combinations of those particular features. For example, where a particular feature is disclosed in the context of a particular aspect, that feature can also be used, to the extent possible, in the context of other aspects.

[0047] Also, when reference is made in this application to a method having two or more defined steps or operations, the defined steps or operations can be carried out in any order or simultaneously, unless the context excludes those possibilities.

[0048] Although specific aspects of this disclosure have been illustrated and described for purposes of illustration, it will be understood that various modifications may be made without departing from the spirit and scope of the invention. Accordingly, the invention should not be limited except as by the appended claims.

[0049]

Claims

WHAT IS CLAIMED IS:

1. An electrodermal activity (EDA) sensor, comprising: one or more pairs of permeable electrodes, configured to be located on skin of a user; a voltage source operable to apply a voltage across the electrodes; and a port electrically coupled to one or more pairs of electrodes to transmit readings from the one or more pairs of permeable electrodes.

2. The EDA sensor as claimed in claim 1, wherein the permeable electrodes comprise one of a spiral electrode, a micro-lace electrode, or a carbon fiber electrode.

3. The EDA sensor as claimed in claim 1, wherein the permeable electrodes comprise a micro-lace electrode having a pitch of one of 50 micrometers, 100 micrometers, or 200 micrometers.

4. The EDA sensor as claimed in claim 1, further comprising a microfluidic channel between the electrodes in each pair of electrodes configured to measure local sweat rate.

5. The EDA sensor as claimed in claim 1, wherein the permeable electrodes comprise a repeating pattern of a geometric shape.

6. The EDA sensor as claimed in claim 1, wherein the voltage source comprises a battery.

7. The EDA sensor as claimed in claim 1, further comprising a microcontroller connected to the port to receive readings from the one or more pairs of electrodes.

8. The EDA sensor as claimed in claim 7, further comprising a wireless module connected to the microcontroller to transmit the readings from the electrodes.

9. The EDA sensor as claimed in claim 8, wherein the wireless module transmits signals according to BlueTooth Low Energy.

10. The EDA sensor as claimed in claim 1, wherein the one or more pairs of electrodes comprise multiple pairs of electrodes and further comprising a multiplexer to multiplex signals from each pair of the multiple pairs of electrodes.

11. The EDA sensors as claimed in claim 1, further comprising an input amplifier to receive voltages across the electrodes as an input.

12. The EDA sensor as claimed in claim 11, further comprising a low pass filter connected to an output of the input amplifier.

13. An electrode permeable to sweat on human skin comprising one of a micro-lace electrode, a spiral electrode, or a carbon fiber electrode.

14. The electrode as claimed in claim 13, wherein the permeable electrodes comprise a micro-lace electrode having a pitch of one of 50 micrometers, 100 micrometers, or 200 micrometers.

15. The electrode as claimed in claim 13, wherein the permeable electrodes comprise a repeating pattern of a geometric shape.

Images