Fig 3 - available from: Scientific Data
This content is subject to copyright. Terms and conditions apply.
Probability density functions (PDF) of physiological signals before and after quality control for the key treatments. Signal data in the perinasal perspiration channel and in the breathing rate channel have passed QC1. Signal data in the EDA, Chest HR, and Wrist HR channels have passed QC 1 and QC2. Hyphenated lines indicate the corresponding distribution means.
Source publication
We describe a controlled experiment, aiming to study productivity and stress effects of email interruptions and activity interactions in the modern office. The measurement set includes multimodal data for n = 63 knowledge workers who volunteered for this experiment and were randomly assigned into four groups: (G1/G2) Batch email interruptions with/...
Contexts in source publication
Context 1
... variables. We screened participants psychometrically via three inventories. Across the sub-scales of these inventories participant distributions feature a healthy spread (Supplemental Fig. S3), suggesting the presence of useful variability. Specifically: Of particular interest in this experiment are conscientiousness, extraversion, and neuroticism traits. Indeed, organizational skills may play some role in the way people manage multi-tasking, while different extraversion and neuroticism levels may affect participant ...
Context 2
... Perceived Stress Scale or PSS -Supplemental Fig. S3c. It measures how stressful the respondents find their lives. The PSS distribution is at 17.13 ± 5.68 in a scale that can range from 0 to 40; thus, it is centered toward the middle of the scale and characterized by absence of extremes that would have confounded physiological ...
Context 3
... of quality control outcomes on physiological variables. Figure 3 depicts the probability density functions (PDF) for each physiological channel before and after quality control; the key treatments are included in each graph. In certain cases, the beneficial effect of quality control is evident. ...
Context 4
... prior to precise analysis, physiological measurements need to be normalized within participants to factor out inter-individual variability, the non-normalized PDF trends in the right column panels of Fig. 3 can provide soundness cues. Indeed, these trends need to conform to common-sense expectations with respect to the effects of the experimental design. In the current experiment, the mean arousal level in the resting baseline (RB) is expected to be the lowest, the mean arousal level in the presentation (PR) is expected to be the highest, ...
Context 5
... validation of study variables. Initial soundness indications for each physiological channel suggested by aggregate and non-normalized PDF trends (Fig. 3), need to be rigorously checked through validity testing. We choose to perform such validation within group and within treatment to account for the nuances of the experimental ...
Citations
... Thermogenic activation, psychogenic activation, and other types of reflex activation of sweat glands have numerous effects on the skin surface (Schwarck et al., 2019;Machado-Moreira and Taylor, 2017;Klous et al., 2020); these are mainly chemical, electrical, and thermal in nature, such as the galvanic skin response (Shahani et al., 1984;Shastri et al., 2012) and the dynamics of the skin temperature distribution (Vainer, 2005;Shilco et al., 2019;Acharya et al., 2014;Bhowmik et al., 2013). The use of thermal imaging cameras with high spatial, temporal, and temperature resolutions has made it possible to visualize not only the temporal dynamics but also the spatial dynamics of sweat gland activity and to describe various associated physiological effects (Vainer, 2005;Ivanitsky, 2006;Shastri et al., 2009;Krzywicki et al., 2014;Zaman et al., 2019;Cardone et al., 2015). The most widely used methods for spatial analysis of thermograms are the matched filter technique (Krzywicki et al., 2014;Familoni et al., 2016), morphological and wavelet analysis (Shastri et al., 2012;Zaman et al., 2019), which are used to delineate individual sweat pores and count them. ...
... The use of thermal imaging cameras with high spatial, temporal, and temperature resolutions has made it possible to visualize not only the temporal dynamics but also the spatial dynamics of sweat gland activity and to describe various associated physiological effects (Vainer, 2005;Ivanitsky, 2006;Shastri et al., 2009;Krzywicki et al., 2014;Zaman et al., 2019;Cardone et al., 2015). The most widely used methods for spatial analysis of thermograms are the matched filter technique (Krzywicki et al., 2014;Familoni et al., 2016), morphological and wavelet analysis (Shastri et al., 2012;Zaman et al., 2019), which are used to delineate individual sweat pores and count them. ...
Sweat gland activity and peripheral hemodynamics, which characterize the function of sympathetic cholinergic nerve fibers and the manifestation mechanisms of vascular tone regulation, respectively, can be detected via dynamic thermography of the skin. Thus, they are useful parameters for diagnosing various forms of neuropathy and functional circulatory disorders. Both parameters affect the dynamics of the skin temperature; therefore, for an adequate description of thermographic data, it is necessary to build models that consider both these coexisting components simultaneously.
Purpose
The objective of this study was to determine the spatiotemporal and statistical features of dynamic thermograms of skin areas with sweat glands and to develop methods for the extraction of temperature components mediated by sweat gland activity (Tsweat) separately from hemodynamics (Tblood) based on thermograms of high and low temperature resolutions.
Methods
To separate the Tsweat and Tblood components, simultaneous thermographic and photoplethysmographic (PPG) measurements were performed in the area of the fingers during a deep inspiratory gasp (DIG). PPG data, which were obtained solely by hemodynamics, were converted into a temperature signal (Tblood) using the spectral filtering approach. By calculating the difference between the skin (Tskin) and blood (Tblood) temperature components, the Tsweat component was determined, which characterizes sweat gland activity and the integrity of the cholinergic sympathetic nerve fibers that innervate them. The Tsweat component was compared with the active sweat pore count curve, which was determined by the adaptive detection of local temperature minima. Thermographic and PPG measurements were performed for 3 min on a group of 15 volunteers during the DIG test. The skin temperature was measured using a cooled thermal imaging camera in the spectral range of 8–9 μm with a temperature sensitivity of 0.02 °C. PPG measurements were performed using a reflectance sensor with a central wavelength of 800 nm. Wavelet analysis with the Morlet basis function was used to preliminarily determine the spectrum of spontaneous temperature oscillations in an area of the skin with and without sweat pores. Statistical parameters of the histogram, such as the standard deviation and the statistical pore activation index (SPAI) - which is proposed in this paper were used in the DIG test to detect sweat gland activity with low-temperature resolution thermograms. The temporal dynamics of the statistical parameters were compared with the dynamics of the sweat pore count.
Results
The Tsweat component was correlated with the sweat pore count on the thermogram with a coefficient of 0.75, confirming the dependence of this temperature component on the sweat gland activity and the need for considering this activity in the analysis of the spatiotemporal dynamics of the human skin temperature. The use of the proposed SPAI and standard deviation allows the detection of sweat gland activity even with thermograms of a low temperature resolution. The use of an integrated map of the sweat gland activity will help the specialist to assess the degree of integrity of the innervation of skin areas in a single image. The primary assessment of the spectrum of temperature oscillations at rest indicated that spontaneous sweat gland activity is accompanied by high-frequency oscillations in the skin temperature localized in the area of sweat pores, within the frequency range of 0.07–0.3 Hz. This suggests the possibility of spectral separation of the temperature component mediated by sweat gland activity from the hemodynamic component, which dominates in the region of <0.1 Hz. The proposed two-component approach for the analysis of skin temperature dynamics allows separate assessment of sympathetic innervation and rhythms of hemodynamic regulation using dynamic thermograms.
There is increasing interest in deploying building-scale, general-purpose, and high-fidelity sensing to drive emerging smart building applications. However, the real-world deployment of such systems is challenging due to the lack of system and architectural support. Most existing sensing systems are purpose-built, consisting of hardware that senses a limited set of environmental facets, typically at low fidelity and for short-term deployment. Furthermore, prior systems with high-fidelity sensing and machine learning fail to scale effectively and have fewer primitives, if any, for privacy and security. For these reasons, IoT deployments in buildings are generally short-lived or done as a proof of concept. We present the design of Mites, a scalable end-to-end hardware-software system for supporting and managing distributed general-purpose sensors in buildings. Our design includes robust primitives for privacy and security, essential features for scalable data management, as well as machine learning to support diverse applications in buildings. We deployed our Mites system and 314 Mites devices in Tata Consultancy Services (TCS) Hall at Carnegie Mellon University (CMU), a fully occupied, five-story university building. We present a set of comprehensive evaluations of our system using a series of microbenchmarks and end-to-end evaluations to show how we achieved our stated design goals. We include five proof-of-concept applications to demonstrate the extensibility of the Mites system to support compelling IoT applications. Finally, we discuss the real-world challenges we faced and the lessons we learned over the five-year journey of our stack's iterative design, development, and deployment.
Antecedents of technology acceptance (TA) are known to be positively associated with measures such as usage intention, behavioral intention, attitude, and satisfaction. Although technology acceptance is investigated widely in prior research, it is not currently clear which variables or factors drive technology acceptance and under different service contexts or conditions. To examine the strength these effects in the artificial intelligence literature, we adopt a meta-analysis approach. We have scoped the literature on artificial intelligence, acceptance measures, and factors affecting acceptance in extant literature. We narrowed our search to business context to find AI-based tools that users, consumers, and customers interact with transactionally, such as chatbots. Findings show AI-based technology factors affect acceptance differently in various service industry contexts as preliminary results. These results have critical implications for researchers and practitioners studying which type of AI-based technology strengthen consumers use in different service contexts. These preliminary findings will be extended to look at interactive relationships of factors affecting acceptance in different contexts.KeywordsTechnology acceptanceArtificial intelligenceMeta-analysisAI factors
Over the past decade, the demand for high-performing knowledge workers (KWs) has grown at an unprecedented rate and shows no signs of slowing. Researchers, designers, engineers, and executives are examples of KWs that perform non-routine, creative work. The work outcomes of KWs as individuals, teams, and organizations play a vital role in the global economy and quality of life. One of the most significant challenges KWs face is balancing stressors on their cognitive and emotional well-being while seeking high productivity. Human cognitive enhancement proposes improving human abilities to acquire and generate knowledge and understand the world. Our cognitive enhancement application for KWs, called the Flow Choice Architecture (FCA), senses their cognitive and affective states, adds context, and recommends appropriate nudges to maximize their healthy flow time. This study provides insights into how FCA implements Human-Centered Design and Responsible Artificial Intelligence (RAI) principles as an interactive AI-powered application that promotes healthy flow performance during knowledge work. FCA applied the RAI tools from Microsoft’s Human-AI eXperience Toolkit to evaluate FCA-specific scenarios. By defining FCA as a hybrid recommendation system and conversational AI agent, we found the following categories of human-AI failure scenarios in FCA: input errors, trigger errors, delimiter errors, and response generation errors. We recommend simulating these errors and undesirable behaviors to improve the design of explainable nudges, meaningful metrics, and well-tuned triggers. The outcome of this RAI evaluation was a robust FCA system design that meets the needs of KWs and enhances their capability to thrive and flourish at work.
KeywordsKnowledge workerCognitive enhancementResponsible artificial intelligence
Anxiety and stress are common emotional responses for human beings, but their chronic manifestation can lead to physical and psychological illnesses. The advancement of sensing technologies, such as Internet of Things, has contributed to the understanding and assisting events related to anxiety and stress. However, the main challenge is knowing which approaches can be used to better monitor these emotions and assist people. Based on a systematic literature review, this work analyzed studies both to determine how data is collected, and to monitor anxiety and stress levels. Two taxonomies synthesize the techniques mapped. The results indicated more emphasis on studying stress than anxiety and more focus on detecting anxiety and stress levels than on assisting the user. Among the main techniques to collect data, 62.5% of the studies used physiological data like heart data, and for data analysis techniques, 48% of the studies used Decision Trees.
The effect of sympathetic activation of sweat glands on the spatio-temporal distribution of skin temperature was studied. The practical significance of the study of sweat glands is supported by the fact that each sweat gland is controlled by the sympathetic cholinergic nerve. Dysfunction of the sympathetic nerve leads to inhibition of the activity of the sweat glands and accompanies many common diseases. In this study, the statistical and spectral properties of the dynamic infrared thermogram of skin with eccrine sweat glands during deep inspiratory gasp (breath test) were investigated. The number of open sweat pores was used as a ground truth measure for the activity of the sweat glands. Results: We found that the activation of the sweat glands, which characterizes the enhancement of the sympathetic response, leads to a change in the statistical spatial distribution of temperatures on the surface of the human body. The histogram of temperature showed a shift to the low-temperature region with a simultaneous increase in the standard deviation. The possibility of using the standard deviation and the skewness (asymmetry of the histogram) to assess the dynamics of the activity of sweat glands was shown. For the first time, the Statistical Pore Activation Index (SPAI) was introduced, which is more adaptive and shows greater accuracy than skewness. The activity of the sweat glands resulted in the appearance of spectral components within the region near 0.1 Hz, which is not typical for oscillations in skin temperature caused by hemodynamics. The results can be used for spectral separation of thermographic data from components caused by hemodynamics and activity of the sweat glands (vasomotor and sudomotor activities). Filtering the dynamic thermograms based on the investigated spectral and statistical features might serve as a method of mathematical processing for mapping the sweat glands and blood vessels on the skin surface. This might be used as a new diagnostic tool in the field of physiology and medicine.
At present, interfaces between users and smart devices such as smart phones rely primarily on passwords. This has allowed for the intrusion and perturbation of the interface between the user and the device and has compromised security. Recently, Frank et al. have suggested that security could be improved by having an interface with biometric features of finger swiping. This approach has been termed touchalytics, in maintaining cybersecurity. The number of features of finger swiping have been large (32) and have been made available as a public database, which we utilize in our study. However, it has not been shown which of these features uniquely identify a particular user. In this paper, we study whether a subset of features that embody human cognitive motor features can be used to identify a particular user. We consider how the security might be made more efficient embodying Principal Component Analysis (PCA) into the interface, which has the potential of reducing the features utilized in the identification of intruders. We compare the accuracy and performance of the reduced feature space to that of having all the features. Embodying a robust continuous authentication system will give users an extra layer of security and an increased sense of peace of mind if their devices are lost or stolen. Consequently, such improvements may prevent access to sensitive information and thus will save businesses money. Consequently, such improvements may prevent access to sensitive information and thus will save businesses money. If continuous authentication models become successful and easily implementable, embodiment and co-adaptation of user authentication would inhibit the growing problem of mobile device theft.
