A Systematic Review and Implementation of Iot-Based Pervasive Sensor-Enabled Tracking System for Dementia Patients

Journal of Medical Systems (2019) 43: 287 https://doi.org/10.1007/s10916-019-1417-z

MOBILE & WIRELESS HEALTH

A Systematic Review and Implementation of IoT-Based Pervasive Sensor-Enabled Tracking System for Dementia Patients

Partha Pratim Ray1 & Dinesh Dash2 & Debashis De3

Received: 30 April 2019 /Accepted: 8 July 2019 /Published online: 17 July 2019 # Springer Science+Business Media, LLC, part of Springer Nature 2019

Abstract In today’s world, 46.8 million people suffer from brain related diseases. Dementia is most prevalent of all. In general scenario, a dementia patient lacks proper guidance in searching out the way to return back at his/her home. Thus, increasing the risk of getting damaged at individual-health level. Therefore, it is important to track their movement in more sophisticated manner as possible. With emergence of wearables, GPS sensors and Internet of Things (IoT), such devices have become available in public domain. Smartphone apps support caregiver to locate the dementia patients in real-time. RF, GSM, 3G, Wi-Fi and 4G technology fill the communication gap between patient and caregiver to bring them closer. In this paper, we incorporated 7 most popular wearables for investigation to seek appropriateness for dementia tracking in recent times in systematic manners. We performed an in-depth review of these wearables as per the cost, technology wise and application wise characteristics. A case novel study i.e. IoT-based Force Sensor Resistance enabled System-FSRIoT, has been proposed and implemented to validate the effectiveness of IoT in the domain of smarter dementia patient tracking in wearable form factor. The results show promising aspect of a whole new notion to leverage efficient assistive physio-medical healthcare to the dementia patients and the affected family members to reduce life risks and achieve a better social life.

Keywords Dementia . Tracking . Sensor system . Wearables . IoT

Introduction Several key symptoms are found among the dementia patients. Sudden memory loss, wandering with time/place, poor Dementia refers to the set of symptoms that cause sudden loss judgement, issues in solving problems and difficulty in com- of memory and mental abilities among patients. It is mainly pleting familiar tasks are to name a few [2]. According to caused by sudden damage and improper functioning of brain Alzheimer’s association (i.e. alz.org), these symptoms are cells while resulting into improper functioning of coordination currently incurable but mostly manageable with proper and efficient communication among several brain-lobes [1]. medication and behavioral care. Around 46.8 million people Such prevalence of insufficient signaling among brain-cells are currently affected by various types of dementia in five hamper regular livelihood activities such as: judgement, think- main continents [3]. The number is expected to reach 74.7 ing, movement, feelings and remembering an incident. million and 131.5 million in 2030 and 2050, respectively. Developing nations of Asia are currently at the highest risk This article is part of the Topical Collection on Mobile & Wireless Health prone zone on the chart whereas African people are the least suffers. The graph in Fig. 1 shows these facts among Africa, * Partha Pratim Ray Europe, Americas (north and south) and Asia. [email protected] Till now, several medical causes have been identified which in turn help into the assimilation of dementia. ’ 1 Department of Computer Applications, Sikkim University, Alzheimer s is the most common cause of dementia which is Gangtok, India observed among the 60–80% of registered cases in the world. ’ 2 Department of Computer Science and Engineering, NIT Patna, Parkinson s disease is another serious cause of dementia, upon Patna, India whose occurrence, affected patient suffers from normal move- – 3 Department of Computer Science and Engineering, MAKAUT, ment related issues [4 7]. Patients having vascular disease, are Kolkata, India found to be lingering with decision making, planning or 287 Page 2 of 21 J Med Syst (2019) 43: 287

come out with novel products to track dementia patients. This article provides an overview of some of these devices being offered to keep track of dementia patients. All of the wearable devices discussed in this study are either currently in market or will be launched very soon. All of the information on device’s features are inherited from respective manufacturer and currently available on their website. This survey is not to be meant for any sorts of device specific analysis. As such, no critical discussion is included of how each manufacturer was effective while development of respective tracking device. The aim of this study is to review existing wearables that help Fig. 1 Prevalence of dementia patients around continents [3] dementia patients to track back to own home by him/her self and by caregivers through in-depth and comprehensive review organizing any event. In mixed dementia cases, sudden mem- of all aspects of surveyed wearables [13–16]. ory loss acts as addendum to the symptoms of the vascular The main contributions of this paper are as follows: inappropriateness. In contrary to this, frontotemporal dementia patients are often found to be more susceptive toward get- & To systematically review the most popular smart wearable ting affected with patientality and behavioral changes. Patients tracking devices for dementia patients’ location tracking; who suffer from cause like Dementia with Lewy Bodies & To perform an in-depth comparative study of the surveyed (DLB), are more prone toward sleep disturbances, visual hal- wearable devices in terms of cost, application, supportive- lucinations, slowness and movement. On the other hand, ness, user experience and technological enrichment; Normal Pressure Hydrocephalus (NPH) compels the patient & To propose a novel, low power consuming and cost- to suffer from the walking disability and uncontrolled. effective IoT-based wearable sensor system i.e. FSRIoT Similarly, the Huntington’s disease takes place due to the ab- to validate the comparative analysis toward catering the normality in a single defective gene, while resulting into the needs of efficient dementia patient tracking in real-life e- occurrence of irritability, depression and mood change among healthcare scenario. the victims. The Wernicke-Korsakoff syndrome (WK) is a & To discusses current technological gap in the surveyed wear- special cause of dementia that is occasionally seen among ables for dementia patient tracking and provide recommen- the alcohol-misusers. Finally, in Creutzfeldt-Jakob Disease dation to further uplift the quality of service in such cases. (CJD), patients get affected with fatal brain disorders. Figure 2 shows different causes of dementia as discusses Rest of the paper is organized as follows. Section II [8–10]. presents the preface to tracking devices. Section III rep- As comprehended from above, three most prevalent and resents the feature description of tracking devices for major challenges for the dementia patients are found as: ab- dementia. Section IV elaborates the user interfaces and normal wandering, sudden memory loss and severe fluctua- services in dementia tracking devices. Section V pre- tion in behavior. Generally, such patients forget to perform sents the various tracking techniques used in the sur- regular livelihood-activities, thus attracting risks into their veyed popular tracking devices. Section VI propose lives. There has been much interest towards development and implements the FSRIoT system model as a proof- and dissemination of relevant technology that can tackle de- of-concept case study to validate the cost-effective mentia patients. Most commonly, tracking solutions for de- tracking device development for dementia patients. mentia patients have been worked upon [11, 12]. Existence Section VII discusses the important technology gaps of these technologies have enabled wearable market players to

Fig. 2 Various causes of dementia Various Causes of Dementia

DLB Vascular Alzheimer’s Parkinson's CJD Disease Disease Mixed NPH

Huntington's Fronto- temporal WK Syndrome J Med Syst (2019) 43: 287 Page 3 of 21 287 and future recommendation to improve current state of Table 1 Wearable device pricing and availability dementia tracking devices. Section VIII concludes the Device Price* Availability article. GPS Smart Sole US$299.00 ξ Available now Freedom GPS Locator Watch US$599.99ϕ Available now Related works to tracking devices Safelink US$149.99 Available now US$199.99ψ ¥ Consumers have shown growing interest in e-health and Mindme Locate US$199.99 Available now – MX-LOCare™ US$129.99 Available now telecare services in recent past [17 31]. Advancement in € fields of Internet of Things (IoT), smart body sensors iTraq3 US$129.00 Pre-order £ and cloud platforms, has embarked into creation of a PocketFinder+ US$159.00 Available now – ξ wearables industry [32 39]. This industry has witnessed *Prices are retrieved from each manufacturer’s web site, US$29.95/ uplift in terms of offering from relatively new comer moth monitoring service extra, ϕ US$10.00 or 35.00 apply per month ψ such as Fitbit, alongside renowned companies like as airtime charge, GS-TRACK price, SL-12 price, ¥ monthly sub- € £ Apple, Sony, Motorola and Garmin [40–51]. Wearable scription US$20.00, monthly subscription US$5.90, monthly subscription US$12.95 products from these vendors allow consumers to gain knowledge of his/her health and fitness status in real- time. While, most of these companies are devoted to as: GPS Smart Sole [53], Freedom GPS Locator production of wearable fitness products, a subset is Watchp [54], Safelink [55], Mindme Locate [56], MX- now engaged at development of dementia tracking de- LOCareTM [57], iTraq3 [58], and PocketFinder+ [59], vice rather than physical monitoring. Main target of have been included in this article for further discussion such devices is to aware the caregivers about patient’s over GPS enabled dementia tracking. Some devices dynamic movement. Global Positioning System (GPS) such as PLI-1000 Patiental Locator System [60], technique is a well-known technique that provides the Revolutionary Tracker [61] and Clevercare are excluded location aware services in technology domain [45–52]. from this study due to their questions about manual Focus of this survey is primarily placed on GPS based antenna-based tracking, generic approach or unavailabil- wireless tracking of dementia patients. Wearable devices ity. Revolutionary Tracker is modelled on top of the that utilize GPS as location tracking module and use Samsung S2 and S3 which proves that it is not a local cellular network (e.g. GSM, CDMA, WCDMA, standalone system. Hence, it is kept out of this study. 4G LTE/VoLTE etc.) for message passing, have been The Clever care has permanently stopped its service and kept aside of this study. Finally, seven devices, such branding, thus not included in this survey.

Fig. 3 The placement of GPS enabled wearable tracking devices on patient’s body. GS- Track, Mindme Locate, PocketFinder+ and iTraq3 and can be placed either in Pendant/ Pocket/Bag/Keyfob/Belt pouch

Chest/Waist/Neck: Pocket, Belt pouch, Pendant GS-Track Mindme Locate Wrist: Watch

iTraq3 Freedom GPS Locator Watch

PocketFinder+ SL-12 MX-LOCareTM

Leg: Shoe GPS Smart Sole 287 Page 4 of 21 J Med Syst (2019) 43: 287

Fig. 4 a GPS Smart Sole placed over inductive charging device. b The makeup of GPS Smart Sole. (Courtesy: GTX Corp SmartSole)

Prototypes published either in research article or exhibited patient’s body where these devices could be worn. There are in science exhibition/fair [62] are also summarily omitted some physical features that the manufacturers have incorpo- from presented survey. Table 1 provides a list of the GPS rated with these wearable devices. GPS Smart Sole, MX- tracking wearable devices covered herein, along with their LOCareTM and PocketFinder+ cases are water resistant. market price and the current status of their availability. GPS Smart Sole (see Fig. 4) and MX-LOCareTM use another fascinating feature i.e. inductive or magnetic charging [63]. Mindme Locate contains multi-network SIM activation feature along with its 48 h battery life. Safelink Feature description of tracking devices includes a SOS button on the device and auto shuts down when power is less than 10%. Notification about GPS has been the center point while researching about how to such event is given to user and caregiver. Freedom GPS track dementia patients. All the devices outlined in this survey Locator (see Fig. 5) Watch has 30 days battery life. It can be categorized as wearables that could be worn regularly has high-impact polycarbonate casing that resists chance by dementia patients throughout a day. Some are designed to of break when accidentally fallen from user’swrist. be worn around wrist as watch (Freedom GPS Locator Watch, iTraq3 has a global SIM card that is accessible to any SL-12, MX-LOCareTM), whereas others could be worn in- network on the globe. Its battery life is of maximum side the shoe (GPS Smart Sole) or in/around Pendant/Pocket/ period than all i.e. 4 months. Motion detection capabil- Bag/Keyfob/Belt pouch (GS-Track, Mindme Locate, iTraq3, ity is another important feature of iTraq3. Table 2 PocketFinder+). Figure 3 shows different locations on shows a longer list of physical features of these devices. Tracking of dementia patients is primarily performed by means of GPS sensors that are pre-embedded into the devices. A GPS sensor periodically receives corresponding latitude and longitude values from the satellites that provides accurate geo-location specifics about the patient. All seven devices, in this study, use GPS sensors as core module into their tracking systems. In addition, GPS Smart Sole includes an altitude sensor [64] to locate the patient’s elevation from mean sea level. MX-LOCareTM (see Fig. 8) is capable to measure number of steps [65] from the starting point of journey. iTraq3 (see Fig. 9) is exception than others that incorporates accelerometer [66] and temperature sensor along with GPS. The temperature senor is positioned at the bottom-layer of the device so that direct skin contact is possible. However, the accelerometer is placed on the side-edge of iTraq3 that can easily sense the body-alignment and movements. Thus, it can regularly check patients’ body temperature and speed of movement whereby providing more safety of a dementia patient. Table 3 presents a detailed list of attached sensors of each Fig. 5 Freedom GPS Locator Watch combo. (Courtesy: Bluewater device. Security Professionals) J Med Syst (2019) 43: 287 Page 5 of 21 287

Table 2 Physical features

Device Features Form factor

GPS Smart Sole Polyurethane sole material, Microcell anti-friction fabric, Wearable (Shoe sole) inductive charging pad, water resistance, battery life: 18–48 h Freedom GPS Locator Watch High-impact polycarbonate case, non-hypoallergenic rubber, Wearable (Watch) 30-days battery life, micro SIM 3FF Safelink SOS button available, not water resistant, auto shut down Wearable when charge is <10% (SL-12:Watch) (GS-Track: PKB*) Mindme Locate 48 h battery life, “drop-in” unit used for easy and fast charging, Wearable (PKB) multi-network SIM activation MX-LOCare™ Magnetic battery charging unit, water proof, battery life: 5 days Wearable (Watch) iTraq3 Global SIM card, battery life: 4 months, Motion detection Wearable (PKB) PocketFinder+ 16 Channel GPS/Glonass module, water resistant Wearable (PKB)

*PKB: Pendant/Pocket/Bag/Keyfob/Belt pouch

User interfaces and services in devices GSM is the mostly used communication technology in all the surveyed products. Mindme Locate and PocketFinder+ None of the wearables included in this study were utilize GPRS as communication media to their web server. customized to be used as a completely stand-alone sys- MX-LOCareTM and PocketFinder+ also use Wi-Fi when tem. These require some sort of companion support for GPS signal strength is very weak (indoor localization). 3G is complete functionality. All of the tracking devices available as an alternative communication technology for discussed here have at least one web portal as feed- users of iTraq3 and PocketFinder+. Freedom GPS Locator back media. Caregivers can track patient’s movement Watch is the only device that uses RF communication when through implied web portals by logging into the sys- short range connectivity is in place. Table 4 describes compat- tem. Mindme Locate (see Fig. 7)provides24×7call ibility, feedback and connectivity means of devices. Geo- center response to its customers. Freedom GPS Locator fencing is a software-based feature that uses GPS and/or Watch leverages a specially designed LCD panel for Radio Frequency Identification (RFID to define a virtual in-system location-aware support for the caregiver. boundary setup around a geographical location. It plays a Caregiver is intimated through alarm and SMS/mail crucial role in localizing activities in surveyed devices [67]. services when the patient is out of the set proximity Existing Android and iOS apps are mostly incorporated with zone. Further, the watch responds to “safezone” and geo-fencing-based web services. Google API could be cus- “live track” requests when out of the geo-fencing tomized in web servers as well as apps for efficient mapping boundary. Safelink (see Fig. 6)isonlydevicethatis of location-aware dispersion [68]. All devices, except sole assisted with web panel services for caregivers. Freedom GPS Locator Watch, deploy geo-fencing services Other devices (GPS Smart Sole, MX-LOCareTM, to users for safer wandering. When a patient trespass in/out iTraq3 and PocketFinder+) are supported by Android through a predefined geofence an SOS alert is served at care- and iOS apps. Family members can keep real-time giver’s avenue either through SMS, Email, Call or Push noti- track over patient’s dynamic behavior by using these fication. Table 5 shows all types of SOS alert and geo-fencing apps. services in these devices.

Table 3 Tracking device sensors Device Type of Sensors

GPS Smart Sole GPS sensor, Altitude sensor Freedom GPS Locator Watch GPS sensor Safelink GPS sensor Mindme Locate GPS sensor MX-LOCare™ GPS sensor, Step count sensor iTraq3 GPS sensor, Accelerometer sensor, Temperature sensor PocketFinder+ GPS sensor 287 Page 6 of 21 J Med Syst (2019) 43: 287

Fig. 8 MX-LOCare™ Watch. (Courtesy: Adiant Mobile) Fig. 6 a SL-12 Watch, b GS-Traq. (Courtesy: Safelink) FSRIoT-A case study Tracking techniques In this paper, we propose a system architecture IoT-based GPSisusedasthemainmethodfortrackingademen- Force Sensor Resistance enabled System i.e. FSRIoT to vali- tia patient in all the devices. Various combination of date the cost effective, wearable and low powered needs for communication technologies has been used to locate the constant monitoring of the dementia or Alzheimer’spa- the patient. GPS Smart Sole requires a 5-min wake up tients. A careful study design, implementation and testing time for initializing the system. Upon successful coordi- were successfully conducted to prove the efficacy of the de- nate’s reception, it sends that information along with ployed scenario. A related implementation was delivered re- speed, bearing and altitude data to GTX monitoring cently [69]. portal. Freedom GPS Locator Watch sends coordinates to caregiver by means of RF communication when user Study design is less than 30 m range of the system. Otherwise, uses GSM to provide necessary information. Safelink uses a The proposed design of the FSRIoT architecture deals with GPS coordinate solution having ±2.2 m accuracy to easy and effective care giving to the dearest elderly member of alert its customers. Mindme Locate intimates the care- family in hassle-free manner. While keeping this objective in giverinevery4mindurationaboutpatient’slocation mind, the FSRIoT architecture is presented in Fig. 11 that with ±10 m accuracy. iTraq3, MX-LOCoreTM and comprises of three major components such as, (i) Patient’s PocketFinder+ uses multi-SIM network to facilitate cus- Sock, (ii) Care Giver’s mobile device and (iii) Cloud tomers with GPS coordinate values. PocketFinder+ (see Service. The experiment was performed in real-life scenario Fig. 10) has fastest location aware services (10 s data where the main author of this paper took part in a homely transmission period) than others. Table 6 lists more detailed step-wise technique for location tracking of devices.

Fig. 7 Mindme Locate. (Courtesy: Mindme) Fig. 9 a iTraq3 Front. b iTraq3 Back. (Courtesy: iTraq Inc.) J Med Syst (2019) 43: 287 Page 7 of 21 287

scenario would certainly act as assistive technology to keep eyes on the beloved aged family member in smart fashion.

Materials used

This study used a set of easily available hardware as well as software components to develop the FSRIoT system. The key items which were effectively used in the experiment are listed in Table 7.

& ATtiny85: This study used an ATtiny85 chip for managing the sensor integration and IoT cloud communication assis- Fig. 10 PocketFinder+. (Courtesy: Location Based Technologies) tance. It is an 8-bit AVR RISC microcontroller having following specifications: 1-SPI, 1-I2C; SRAM- 512bytes; environment. The hardware assembly consisting of microcon- EEPROM-512 bytes; 20 MIPS; 5 PWM Pins; Program troller, ESP-07 module and battery were tied together with a memory-8kB. It was used in this aspect due to its small daily used sock which was connected with the generic Force size and low power consuming behavior. Sensing Resistor (FSR) attached at the toe of leg, as shown in & Resistance: A generic 3.3KOhm, 1/4 W, 5% Tolerance Fig. 11. The experiment was simulated in such way so that the resistor was used to act as voltage divider with respect to sock was associated with the leg of a real dementia patient the analog signal received from the FSR sensor. who is ready to roam around in and around his/her house. The & FSR Sensor: A popular version of FSR sensor was de- test was performed during morning session between 7.00– ployed in this regard that had following specifications. 7.45 AM. The goal of this study was to check whether pro- Semi-conductive Layer-0.05 in.; Spacer/Rear Adhesive- posed architecture can be associated with one of the popular 0.006/0.002 in.; Acrylic Conductive Layer-0.005 in.. IoT cloud services for better analysis and prompt caregiving. The FSR sensor is built using the Polymer Thick Film Whenever, the sock-worn simulated patient moved FSR sen- (PTF) material which provides a nice small sensing area sor got pressed and thus internal resistance changed which with flexible electrodes. The flexibility of the electrodes indirectly changed the voltage level. This analog signal was used to attach it on the toe of the sock. The structure of the then processed by the microcontroller and was instantly sent FSR sensor is shown in Fig. 12. to the nearby caregiver’s mobile device in form of visual no- & ESP-07: It is a variant of the popular ESP8266 series that tification on the web page hosted at the ESP-07 module. comes in small and compact size. The specifications are as Otherwise, when the caregiver is out of station, he/she could follows, such as: ISM 2.4 GHz; IEEE 802.11 b/g/n continuously monitor the real-time movement-status of the Standard Protocol; 3.3–3.7 V, and Current 40-160 mA. dementia patient who is currently residing at the home. This The power consumption is also very minimum within

Fig. 11 Proposed architecture- FSRIoT for FSR to User communication Cloud Service Microcontroller, ESP-07 and Battery Assembly Internet Wake Up! Walk! FSR Sensor Amazed! Astonished! Sleep!

Patient’s Sock Care Giver 287 Page 8 of 21 J Med Syst (2019) 43: 287

electrode of FSR was extended to provide the voltage divider circuit. The analog read value was sent to the Pin 2 of ATtiny85. Pin 5 and 7 of ESP-07 were used as Rx and Tx, respectively. The connections between ATtiny85 and ESP- Flexible Substrate with Printed Semi-Conductor 07 are herein presented vividly. The 3.7 V Li-Po battery was Spacer connected to all the elements of the proposed schematic. Opening & Algorithm and Flow Chart: The flow chart and algorithm of the FSRIoT are presented in Fig. 14 and Algorithm 1,2 and 3, Spacer Adhesive respectively. The flow chart of the system starts with Vent powering ON the system. The baud rate is fixed within range of 9600–115200 bps per requirement of specific usage pol- icy. Calibration of FSR sensor was done to check if it ready Sensing Area Flexible PTF Electrodes to serve. Later, two types of configurations namely client and Tail server are formulated. When configured as client, the system Fig. 12 Structure of the FSR sensor [70, 71] senses the force value and sends it to the remote IoT cloud for data storage and analysis purposes. Else if, it is conFig. d as server, nearby caregiver can directly access the real-time 12-50 mW. It could be conFig. d either as client, server or foot-steps over the hosted webpage via the mobile device. repeater in local network. Actually, the services are developed such way that the care- & Li-Po Battery: A 3.7 V, 150mAh capacity Lithium- giver can have real-time access on the current status of move- Polymer battery was used to power the system. ment whether placed in-house or in distant location. The & Wires and miscellaneous: Few Male-Male/Female wires and algorithm for FSRIoT access and configuration are illustrated some passive electrical items were used to attach the system. inListing1.Ithastwoofthreefunctionsreadytoworkatany & Arduino IDE: In proposed study, version 1.8.7 was used to point of time. In both of the cases, FSR_Sense_Monitor() is code, compiler, run and deploy the algorithm. called to execute. This necessary, otherwise force sensor data could not sent and processed. The force or resistance conver- sion formulas are opted from the FSR datasheet and openly distributed public e-learning domains. Methodology

The development phase of the FSRIoT consists of two items Results and discussion (i) embedded sensor assembly and (ii) algorithm. The FSRIoT architecture was tested and evaluated a number & Sensor Assembly: The proposed schematic is illustrated in of times. Although, due to obligations of ethical issues, none Fig. 13 where all Pin-wise connections are shown. One of the local dementia patients was available to perform this

FSR 0.5"

ESP-07 +- 3.3V

3.3V 1 16 Pin 7 15 Pin 5 87 65 Li-Po 3 ATtiny 85 Battery Tttiny85 ESP-07 ESP8266 1234 3.3V 3.3K 8 9 R1 FSR 0.5" Pin 2 ATtiny 85 Fig. 13 Proposed schematic of the proposed FSR sensor deployment J Med Syst (2019) 43: 287 Page 9 of 21 287

Fig. 14 Flow chart of the Start working of the proposed system Power On System

Set Baud Rate @ 9600- 115200 bps

Calibrate FSR

Is Client Measure FSR Yes Configured ?

Is No No Pressure > Is Threashold ? Server Yes Measure FSR Configured ? Yes

Send Current No Is Movement Pressure > No Is Status-Data to Threashold ? Remote IoT Time/ Cloud Connection/ Battery Yes Out ? Real-time Send Current Access and/or Yes Movement Analyze by Status-Data to Stop Client the Locally Connected WiFi enabled Maintain Time Devices and Date of Such Movement Visualize Movement Time and Date in Real-Time

experiment over them, the test seemed to successful per the patients upon their agreements and ethical clearance. Further, obtained results. The deployment was in prototypical struc- battery life calculation and wearable-pervasive prototypes ture, thus leaving opportunity to further improve it per more could be provisioned. Table 7 presents the comparisons be- compact size format in future. Pressure (lbs) and foot-steps vs tween the proposed FSRIoT and the surveyed systems. time are plotted in real-time as shown in Fig. 15aandb. FSRIoT has evolved as the low-cost and open source design Interestingly, pressure plot has a resemblance to the foot- centric sensor system that is IoT ready. Its form factor is suit- steps plot, as both of the data were parallel sent to either to able for dementia patients to wear around leg-toes which the IoT cloud service or locally connected user. The test was makes its credibility higher than other alternatives. performed for 20 times each of specified duration. It was as- sumed that simulated scenario maps with the dementia patient’s real-life movement during morning session that in- Technology gap and future recommendation cludes sitting, amazed, walk and stop postures. Out of 20 times, 100% accuracy was observed for the case of detection Presented wearables are mostly designed upon the geo- of foot-stepping and continuous communication with the ex- fencing concept. Wearables are attached on different body ternal caregiver’s mobile device or remote cloud. However, in parts of the dementia patient and a virtual geographical bound- future the proposed system would be tested against dementia ary is estimated by the caregivers [72–85]. Whenever, the 287 Page 10 of 21 J Med Syst (2019) 43: 287

Fig. 15 a Time vs Pressure (lbs), b Time vs Foot-steps Pressure (lbs)

7.05 AM 7.10 AM 7.15 AM 7.20 AM 7.25 AM 7.30 AM 7.35 AM 7.40 AM Time (a) Foot Steps

7.05 AM 7.10 AM 7.15 AM 7.20 AM 7.25 AM 7.30 AM 7.35 AM 7.40 AM Time (b) dementia patient enters into the “safe zone” or exits out from the presented systems. None of the devices has in- it, several call alerts or SMS are sent over the caregiver’s cluded specific bio-sensors for seamless healthcare smartphone or pre-registered web portal. In all the devices, of the dementia patients when they are out of home. similar facilities are available [86–104]. However, few impor- Pulse rate, SpO2 and Galvanic Skin Response (GSR) tant design specifications are missing in current devices that sensors can provide better e-healthcare services than must be sought for more efficient dementia tracking services, what are being used now [105–110]. as presented below. & Machine learning algorithm prediction: None of the wearables reveals whether it has used any sort of & Lack of bio-sensors: All the wearables are currently machine learning algorithm in predicting the pa- integrated with GPS sensor in their designs. A hand- tient’s movement or wandering pattern. For example, ful of selective sensors, such as: temperature, accel- a dementia patient may feel unwell when outside of erometer, step counter and altitude are equipped with home. Advanced, machine learning algorithm must

Table 4 Device user interfaces Device Compatibility Feedback Connectivity

GPS Smart Sole Android, iOS App, Web portal GSM/2Gξ Freedom GPS Locator Watch Proprietary LCD panel, Web portal GSM, RF Safelink Proprietary Web panel GSM Mindme Locate Proprietary Response center, Web portal GSM, GPRS MX-LOCare™ Android, iOS App, Web portal GSM, Wi-Fi iTraq3 Android, iOS App, Web portal GSM, 3G PocketFinder+ Android, iOS App, Web portal GSM, GPRS, 3G, Next G, Wi-Fi, Cell ID

ξ T-mobile, AT&T network J Med Syst (2019) 43: 287 Page 11 of 21 287

Table 5 Device services Device Geo-fencing service SOS Alert Service

GPS Smart Sole Available SMS, Email Freedom GPS Locator Watch Not available Email, SMS Safelink Available Calling Mindme Locate Available Email, Calling MX-LOCare™ Available Email, SMS iTraq3 Available Push notification PocketFinder+ Available Email, Push notification

in place to at least analysis whether the scenario is dementia patients. Such interventions may get seriously life-threatening or a little rest will make the patient affected when new software-services (technology-stacks) feel better [111–115]. Novel algorithms and light- are provisioned. Specialized cloud service may provide weight IoT-based schemes should be developed to more flexible user-centric supports toward real-time geo- cater the features like “amazedness”, “walk” or fencing setup, tracking of movement history, e-health sta- “still” status of physio-psycho motor attributes of tus etc. [116–118]. the dementia patents. None of the surveyed and the & Power consumption: All of the existing deployments proposed FSRIoT deal with such interventions. are powered by rechargeable-battery packs that are & Flexibility in geo-fencing: Current systems use pre- normally drained out within 1–2daysofoperation deployed geo-fencing services for tracking of the demen- time. The reason is current devices use power hun- tia patients. However, it is possible that a patient may go gry cellular radio spectrum technologies such as outside the predefined geo-boundary for some livelihood GSM, 3G and 4G LTE. Low power wide area pro- needs. In that case, caregivers will be unnecessarily tocols such as: LTE-M, NB-IoT and EC-GSM-IoT interrupted. Hence, more flexible geo-fencing service may be used for longer battery-life, thus enhancing should be in place. the overall tracking service. & Need for specialized cloud service: Most of the present & User experience: The wearables are not persistent with the tracking devices use proprietary web-server for dissemi- user experience and adoption facilities. For example, in nating the real-time visualization service for movement of some devices, a key-press button (SOS button) is included

Table 6 Location tracking technique in devices Device Technique

GPS Smart Sole 5-min GPS wakeup time; Receives GPS coordinates from satellite and sends to GTX monitoring portal for location tracking; Speed, bearing and altitude information are also sent to remote server Freedom GPS Locator Watch Receives coordinates from satellite; Sends location alert to caregiver by RF and GSM services when <30 m (indoor) and > 200 m (outdoor) range from GPS locator and caregiver’s detector node, respectively Safelink Receives coordinates from satellite; Sends location alert to caregiver by GSM with ±2.2 m accuracy; Also used cell towers for location tracking when no enough GPS signal in sight (indoor) Mindme Locate Receives coordinates from satellite; Sends location alert to caregiver by multi-network GSM with ±10 m accuracy in every 4 min; MX-LOCare™ Receives coordinates from satellite (outdoor) and Wi-Fi (indoor); Sends location alert to caregiver by GSM/2G/3G/Next G iTraq3 Receives coordinates from satellite; Sends location alert along with motion and body temperature to caregiver by GSM/3G in every 5 min; Multiple care giver sharing is available PocketFinder+ Receives coordinates from satellite (outdoor) and Wi-Fi (indoor); Sends location alert to caregiver by GSM/GPRS/3G/Cell ID in every 10s; Multiple care giver sharing is available 287 Page 12 of 21 J Med Syst (2019) 43: 287

Table 7 Materials used in this study

Sl. No. Item Manufacturer Specifications

1 ATtiny85 ATMEL 8-bit AVR RISC microcontroller; 1-SPI, 1-I2C; SRAM- 512bytes; EEPROM-512 bytes; 20 MIPS; 5 PWM Pins; Program memory-8kB 2 FSR Sensor Interlink Electronics Semi-conductive Layer-0.05 in.; Spacer/Rear Adhesive-0.006/0.002 in.; Acrylic Conductive Layer-0.005 in.; PTF material 3 Resistance Generic 3.3KOhm, 1/4 W, 5% Tolerance 4 ESP-07 ESP8266 Sunrom Electronics ISM 2.4 GHz; IEEE 802.11 b/g/n Standard Protocol; 3.3–3.7 V, Current 40-160 mA 5 Li-Po Battery Generic 180mAh, 3.7 V 6 Wires Generic Male-Male/Female Jumper Wire 7 Arduino IDE Many Version 1.8.7

for intimating the caregivers about an emergency situa- more feasible for the user of developing nations by making tion. But it is actually not a practical solution for a demen- their cost and recurring service low. Special attention should tia patient. Some more ubiquitous and pervasive approach be given for leveraging of cloud based and machine learning may be incorporated with the systems [119, 120]. assisted support to the low power consuming products. & Cost: It is observed that most of the devices incur monthly Further, it can be prescribed that the proposed FSRIoT, as or yearly subscription charges for providing real-time alternative to other surveyed wearables, is more beneficial to tracking services to caregivers. Dementia patients and the common end-users, dementia patients, and care givers. caregivers who originate from various developing coun- Due to its low-cost design, open source approach, IoT- tries are often deprived of economic freedom. Hence, enablement, and attractive form factor, a large pool of people more cost-effective service is required [121–145]. suffering from dementia and similar other diseases could be & Neuro Imaging: Computer aided algorithm may be envis- facilitated with more efficient tracking service. aged and incorporated within the existing IoT and wearable framework to early diagnose and detect the possibil- ity of dementia [146–180]. Neuro imaging-based ad- Conclusion vanced computational approaches are getting seamless popularity in recent times toward achieving this goal With advancement of accurate GPS monitoring technique, [181–206]. wearable-entrepreneurs have developed a number of tracking devices like GPS Smart Sole, Freedom GPS Locator Watch, Safelink, Mindme Locate, MX-LOCareTM, iTraq3 and Recommendation To this end, it is recommended that users PocketFinder+ to efficiently locate dementia patients. These should choose a dementia tracking device based on the three devices not only track regular wandering and dynamic move- criteria such as: flexibility of geo-fencing service, availability ment of dementia patients but also counts steps, speed, bearing of emergency support and appropriate user experience. and body temperature, thus resulting into safer consumer cen- Manufacturers are hereby suggested to make their designs tric approach of smart living. Proposed FSRIoT is developed,

Table 7 Comparison Between the Surveyed Devices and Proposed FSRIoT Sensor System

Device Design Cost (US$) Sensor IoT Connectivity Form Factor

GPS Smart Sole Proprietary 299.00 GPS No 2G Wearable (Shoe sole) Freedom GPS Locator Watch Proprietary 599.99 GPS No 2G, RF Wearable (Watch) Safelink Proprietary 149.99 GPS No 2G Wearable (PKB*) Mindme Locate Proprietary 199.99 GPS No 2G, GPRS Wearable (PKB) MX-LOCare™ Proprietary 199.99 GPS No 2G, Wi-Fi Wearable (Watch) iTraq3 Proprietary 129.99 GPS No 2G, 3G Wearable (PKB) PocketFinder+ Proprietary 129.00 GPS No 2G, GPRS, 3G, Wi-Fi Wearable (PKB) FSRIoT Open Source 15.00 FSR Yes Wi-Fi, 2G, 3G, 4G LTE Wearable (Sock)

*PKB: Pendant/Pocket/Bag/Keyfob/Belt pouch J Med Syst (2019) 43: 287 Page 13 of 21 287 validated and tested against a simulation environment that has wearables smarter to counter the tracking of the dementia showed a promising direction toward real-time monitoring patients in near future. and tracking of the dementia patients with help of IoT and wearable devices. We also pointed out several key technolog- Compliance with ethical standards ical gaps and future prospects of the reviewed wearables that are supported from IoT-based ecosystems. Major concern Conflict of interest The authors declare no conflict of interest. should be imposed on the machine learning, neuro imaging and power consumption related factors to make the IoT-based

Appendix 1 287 Page 14 of 21 J Med Syst (2019) 43: 287 J Med Syst (2019) 43: 287 Page 15 of 21 287 287 Page 16 of 21 J Med Syst (2019) 43: 287

References 21. Fraser, K. C., Meltzer, J. A., and Rudzicz, F., Linguistic features identify Alzheimer’s disease in narrative speech. J. Alzheimer’s Disease 49(2):407–422, 2015. 1. Alzheimer’s Disease and Dementia. Dementia Types: Signs, 22. Roark, B., Mitchell, M., Hosom, J.-P., Hollingshead, K., and Symptoms, and Diagnosis, 2017. Available: https://www.alz.org/ Kaye, J., Spoken language derived measures for detecting mild dementia/types-of-dementia.asp cognitive impairment. IEEE Trans. Audio, Speech, Language 2. 10 Warning Signs of Alzheimer’s, 2017. Available: https://www. Process. 19(7):2081–2090, Sep. 2011. alz.org/national/documents/tenwarnsigns.pdf 23. Aramaki, E., Shikata, S., Miyabe, M., and Kinoshita, A., 3. World Alzheimer Report 2015 launched, 2015. Available: https:// Vocabulary size in speech may be an early indicator of cognitive ncdalliance.org/news-events/news/world-alzheimer-report-2015- impairment. PLoS One 11(5):e0155195, 2016. launched 24. Wechsler, D., WAIS-III: Administration and Scoring Manual: ’ 4. Living with Alzheimer s, 2015. Available: https://www.wearable- Wechsler Adult Intelligence Scale. San Antonio: Psychological technologies.com/2015/12/living-with-alzheimers/ Corp., 1997. 5. Thapliyal, H., Khalus, V., and Labrado, C., Stress Detection and 25. Tanaka, H., et al., Automated social skills trainer. In Proc. 20th Int. Management: A Survey of Wearable Smart Health Devices. IEEE Conf. Intell. User Interfaces, 2015, pp. 17–27. – Consumer Electronics Magazine. 6(4):64 69, 2017. 26. Jimison, H., Pavel, M., McKanna, J., and Pavel, J., Unobtrusive 6. Ray, P. P., Understanding the Role of Internet of Things Towards monitoring of computer interactions to detect cognitive status in – Smart e-Healthcare Services. Biomed. Res. 28(8):1604 1609, elders. IEEE Trans. Inf. Technol. Biomed. 8(3):248–252, 2004. 2017. 27. Y. Sakai, Y. Nonaka, K. Yasuda, and Y. I. Nakano, Listener agent 7. Ray, P. P., A survey of IoTcloud platforms. Future Computing and for elderly people with dementia. In Proc. 7th Annu. ACM/IEEE – – Informatics Journal. 1(1 2):35 46, 2016. Int. Conf. Hum.-Robot Interact., 2012, pp. 199–200. 8. Seneviratne, S., Hu, Y., Nguyen, T., Lan, G., Khalifa, S., 28. Folstein, M. F., Folstein, S. E., and McHugh, P. R., Mini-mental Thilakarathna, K., Hassan, M., and Seneviratne, A., A Survey of state’. A practical method for grading the cognitive state of pa- Wearable Devices and Challenges. IEEE Communications tients for the clinician. J. Psychiatr. Res. 12(3):189–198, 1975. – Surveys and Tutorials. 19(4):2573 2620, 2017. 29. Harden, R. M., and Gleeson, F., Assessment of clinical compe- 9. Mosenia, A., Sur-Kolay, S., Raghunathan, A., and Jha, N. K., tence using an objective structured clinical examination (OSCE). Wearable Medical Sensor-Based System Design: A Survey. Med. Educ. 13(1):39–54, 1979. IEEE Transactions on Multi-Scale Computing Systems. 3(2): 30. Tanaka, H., Adachi H., Ukita N., Kudo T., and Nakamura S., 124–138, 2017. Automatic detection of very early stage of dementia through mul- 10. Ismail, S., Barton, A., Gaydecki, P., Karimian, N., Karimian, S., timodal interaction with computer avatars. In Proc. 18th ACM Int. Design and development of a compact wearable dipole GPS an- Conf. Multimodal Interact., 2016, pp. 261–265. tenna. In: Proceedings of IEEE 46th European Microwave 31. Orimaye, S. O., Wong, J. S.-M., and Golden, K. J., Learning Conference (EuMC), 2016, p. 96–99. predictive linguistic features for Alzheimer’s disease and related 11. McKhann, G. M. et al., The diagnosis of dementia due to dementias using verbal utterances. In Proc. 1st Workshop Alzheimer’s disease: Recommendations from the National Comput. Linguistics Clin. Psychol. (CLPsych), 2014, pp. 78–87. Institute on aging-Alzheimer’s association workgroups on diag- 32. König, A. et al., Automatic speech analysis for the assessment of nostic guidelines for Alzheimer’s disease. Alzheimer’sDementia patients with predementia and Alzheimer’s disease. Alzheimer’s 7(3):263–269, 2011. Dementia, Diagnosis, Assessment Disease Monitor. 1(1):112– 12. Santacruz, K. S., and Swagerty, D., Early diagnosis of dementia. 124, 2015. Amer. Family Phys. 63(4):703–713, 2001. 33. Kaplan, E., Goodglass, H., and Weintraub, S., Boston Naming 13. Taler, V., and Phillips, N. A., Language performance in Test. Austin: Pro-ed, 2001. Alzheimer’s disease and mild cognitive impairment: A compara- 34. Vinciarelli, A., Pantic, M., and Bourlard, H., Social signal process- tive review. J. Clin. Exp. Neuropsychol. 30(5):501–556, 2008. ing: Survey of an emerging domain. Image Vis. Comput. 27(12): 14. Laske, C. et al., Innovative diagnostic tools for early detection of 1743–1759, 2009. Alzheimer’s disease. Alzheimer’s Dementia 11(5):561–578, 2015. 35. Yang F., Sehili, M. A., Barras, C., and Devillers, L., Smile and 15. Geldmacher, D. S., Cost-effective recognition and diagnosis of laughter detection for elderly people-robot interaction. In Proc. Int. dementia. In: Seminars Neurology. Vol. 22. New York: Thieme Conf. Social Robot., 2015, pp. 694–703. Medical Publishers, 2002, 063–070. 36. Asplund, K., Norberg, A., Adolfsson, R., and Waxman, H. M., 16. Clark, H. H., and Tree, J. E. F., Using uh and um in spontaneous Facial expressions in severely demented patients—Astimulus– speaking. Cognition 84(1):73–111, 2002. response study of four patients with dementia of the Alzheimer – 17. Kato, S., Endo, H., Homma, A., Sakuma, T., and Watanabe, K., type. Int. J. Geriatric Psychiatry 6(8):599 606, 1991. Early detection of cognitive impairment in the elderly based on 37. Sturm, V. E. et al., Mutual gaze in Alzheimer’s disease, Bayesian mining using speech prosody and cerebral blood flow frontotemporal and semantic dementia couples. Soc. Cogn. activation. In Proc. 35th Annu. Int. Conf. IEEE Eng. Med. Biol. Affect. Neurosci. 6(3):359–367, 2010. Soc. (EMBC), 2013, pp. 5813–5816. 38. Shimokawa, A. et al., Influence of deteriorating ability of emo- 18. Pirozzolo, F. J., Christensen, K. J., Ogle, K. M., Hansch, E. C., and tional comprehension on interpatiental behavior in alzheimer-type – Thompson, W. G., Simple and choice reaction time in dementia: dementia. Brain Cogn. 47(3):423 433, 2001. Clinical implications. Neurobiol. Aging 2(2):113–117, 1981. 39. Fernandez-Duque, D., and Black, S. E., Impaired recognition of 19. Jarrold, W. et al., Aided diagnosis of dementia type through com- negative facial emotions in patients with frontotemporal dementia. – puter based analysis of pontaneous speech. Proc. ACL Workshop Neuropsychologia 43(11):1673 1687, 2005. Comput. Linguistics Clin. Psychol.:27–36, 2014. 40. Baylor, A. L., and Kim, Y., Pedagogical agent design: The impact 20. Lehr, M., Prud, E. T., Shafran, H. I., and Roark, B., Fully auto- of agent realism, gender, ethnicity, and instructional role. In mated neuropsychological assessment for detecting mild cognitive Intelligent Tutoring Systems (Lecture Notes in Computer – impairment. Proc. INTERSPEECH:1039–1042, 2012. Science) vol. 3220. Berlin, Germany: Springer, 2004, pp. 592 603. J Med Syst (2019) 43: 287 Page 17 of 21 287

41. Bickmore, T., and Cassell, J., Small talk and conversational story- 65. How Pedometers Work?, 2015. Available: https://www.runtastic. telling in embodied conversational interface agents. In Proc. com/blog/en/hot-topics-and-reviews/how-pedometers-work/ AAAI Fall Symp. Narrative Intell., 1999, pp. 87–92. 66. Accelerometers: Introduction to Acceleration Measurement, 2017. 42. Tanaka, H., Yoshino, K., Sugiyama, K., Nakamura, S., and Available: https://in.omega.com/prodinfo/accelerometers.html Kondo, M., Multimodal interaction data between clinical psychol- 67. What is Geo-Fencing? Putting location to work, 2017. Available: ogists and students for attentive listening modeling. In Proc. Conf. https://www.cio.com/article/2383123/mobile/geo-fencing- Oriental Chapter Int. Committee Coordination Standardization explained.html Speech Databases Assessment Techn. (O-COCOSDA), 2016, 68. Geo-Fencing API, 2017. Available: https://www.w3.org/TR/geo- pp. 95–98. fencing/ 43. Endo, T. et al., Initial response time measurement in eye move- 69. Raza, U., Kulkarni, P., and Sooriyabandara, M., Low Power Wide ment for dementia screening test. In: Proc. 15th IAPR Int. Conf. Area Networks: An Overview. IEEE Communications Surveys & Mach. Vis. Appl. (MVA), 2017, pp. 262–265. Tutorials. 19(2):855–873, 2017. 44. Heldner, M., and Edlund, J., Pauses, gaps and overlaps in conver- 70. Shinozuka, K., Wearable Sensors: A Novel Healthcare Solution – sations. J. Phon. 38(4):555 568, 2010. for the Aging Society. https://archive.googlesciencefair.com/ 45. American Psychiatric Association, Diagnostic and statistical man- projects/en/2014/ ual, 4th edn, text revision (DSM-IVTR). Washington: Amer. d5b1aaea6cc6d609d3cca8ad616d2aa39c06258ad1632 Psychiatric Assoc., 2000. 7cbfaf8250de85d876c 46. Crawford, J. R., and Garthwaite, P. H., Investigation of the single 71. FSR guide. https://www.sparkfun.com/datasheets/Sensors/ case in neuropsychology: Confidence limits on the abnormality of Pressure/fsrguide.pdf test scores and test score differences. Neuropsychologia 40(8): 72. FSR Hook-Up Guide. https://learn.sparkfun.com/tutorials/force- – 1196 1208, 2002. sensitive-resistor-hookup-guide 47. Mower, E., Black, M. P., Flores, E., Williams, M., and Narayanan, 73. Saragih, J. M., Lucey, S., and Cohn, J. F., Face alignment through S., Rachel: Design of an emotionally targeted interactive agent for subspace constrained mean-shifts. In Proc. IEEE 12th Int. Conf. children with autism. In Proc. IEEE Int. Conf. Multimedia Expo Comput. Vis., 2009, pp. 1034–1041. (ICME), 2011, pp. 1–6. 74. I. Naim, M. I. Tanveer, D. Gildea, and Hoque, M. E., 2015. 48. Bishop, C. M., Pattern recognition. In: Machine Learning. Vol. Automated analysis and prediction of job interview performance. 128. New York, NY, USA: Springer-Verlag, 2006, 1–58. Available: https://arxiv.org/abs/1504.03425 49. Hanson, M., Glottal characteristics of female speakers. Ph.D. dis- 75. Lyons, M., Akamatsu, S., Kamachi, M., and Gyoba, J., Coding sertation, Dept. Division Appl. Sci., Harvard Univ., Cambridge, facial expressions with Gabor wavelets. In Proc. 3rd IEEE Int. 1995. Conf. Autom. Face Gesture Recognit., 1998, pp. 200–205. 50. Bucks, R., Singh, S., Cuerden, J. M., and Wilcock, G. K., Analysis 76. Cohn, J. F., et al., Detecting depression from facial actions and of spontaneous, conversational speech in dementia of alzheimer vocal prosody. In Proc. 3rd Int. Conf. Affective Comput. Intell. type: Evaluation of an objective technique for analysing lexical Interact. Workshops (ACII), 2009, pp. 1–7. performance. Aphasiology 14(1):71–91, 2000. 77. Metz, C. E., Basic principles of ROC analysis. Semin. Nucl. Med. 51. Wright, H. H., Silverman, S., and Newhoff, M., Measures of lex- – ical diversity in aphasia. Aphasiology 17(5):443–452, 2003. 8(4):283 298, 1978. 52. Herbert, R., Hickin, J., Howard, D., Osborne, F., and Best, W., Do 78. Lunsford, R., Heeman, P. A., Black, L., and Santen, J. V., Autism ‘ ’ ‘ picturenaming tests provide a valid assessment of lexical retrieval and the use of fillers: Differences between um and uh. In Proc. – in conversation in aphasia? Aphasiology 22(2):184–203, 2008. DiSS-LPSS Joint Workshop, 2010, pp. 107 110. 53. GPS Smart Sole, 2017. Available: http://gpssmartsole.com/ 79. Chaspari, T., Bone, D., Gibson, J., Lee, C.-C., and Narayanan, S., gpssmartsole/ Using physiology and language cues for modeling verbal response 54. Freedom GPS Locator Watch, 2017. Available: http://www. latencies of children with ASD. In Proc. IEEE Int. Conf. Acoust., – bluewatersecurityprofessionals.com/elderlytracking.htm Speech Signal Process. (ICASSP), 2013, pp. 3702 3706. 55. Safelink, 2017. Available: http://safelinkgps.com/ 80. Ray, P. P., Dash, D., and De, D., Edge Computing for Internet of 56. Mindme Locate, 2017. Available: http://www.mindme.care/ Things: A Survey, e-Healthcare Case Study and Future Direction. mindme-locate.html Journal of Network and Computer Applications, Elsevier, 2019, 57. MX-LOCare™, 2017. Available: https://www.adiantmobile.com/ (Accepted) (SCI) (IF: 3.9). mx-locare-watch/ 81. Ray, P. P., Data Analytics: India Needs Agency for Health Data. 58. iTraq3, 2017. Available: https://www.itraq.com/ Curr. Sci. 112(6):1082, 2017. 59. PocketFinder+, 2017. Available: http://pocketfinder.com/ 82. Ray, P. P., Internet of Things Cloud Enabled MISSENARD Index patientaltracker/ Measurement for Indoor Occupants. Measurement, Elsevier 92: 60. PLI-1000 Patiental Locator System, 2017. Available: https:// 157–165, 2016. https://doi.org/10.1016/j.measurement.2016.06. projectlifesaver.org/locating-technology/pli-1000-patiental- 014. locator-system/ 83. Zhou, L., Fraser, K. C., and Rudzicz, F., Speech recognition in 61. Revolutionary Tracker, 2017. Available: https://matt81853. Alzheimer’s disease and in its assessment. Proc. INTERSPEECH: wixsite.com/website 1948–1952, 2016. 62. This Kid Designed A Wearable Health Device For His 84. Meilán, J. J., Martínez-Sánchez, F., Carro, J., López, D. E., Grandfather With Alzheimer’s, 2016. Available: https://www. Millian-Morell, L., and Arana, J. M., Speech in Alzheimer’sdis- fastcompany.com/3034118/this-kid-designed-a-wearable-health- ease: Can temporal and acoustic parameters discriminate demen- device-for-his-grandfather-with-alzheimers tia? Dementia Geriatric Cognit. Disorders 37(5–6):327–334, 63. Wang, Z., Cui, S., Han, S., Song, K., Zhu, C., Matveevich, M. I., 2014. and Yurievich, O. S., A Novel Magnetic Coupling Mechanism for 85. Horley, K., Reid, A., and Burnham, D., Emotional prosody per- Dynamic Wireless Charging System for Electric Vehicles. IEEE ception and production in dementia of the Alzheimer’stype.J. Trans. Veh. Technol.:99, 2017 (pre-print). Speech, Lang., Hearing Res. 53(5):1132–1146, 2010. 64. Sensors for Altitude Measurement, 2017. Available: http://www. 86. Paeschke, A., and Sendlmeier, W. F., Prosodic characteristics of amsys.info/special/height.htm emotional speech: Measurements of fundamental frequency 287 Page 18 of 21 J Med Syst (2019) 43: 287

movements. In Proc. ISCA Tuts. Res. Workshop (ITRW) Speech 105. Doukas, C., and Malogiannis, I., Bringing IoT and cloud comput- Emotion, 2000, pp. 75–80. ing towards pervasive healthcare. Proceedings of IEEE 87. Ray, P. P., A Survey on Visual Programming Languages in International Conference on Innovative Mobile and Internet Internet of Things. Scientific Programming, Hindawi, Vol. 2017, Services in Ubiquitous Computing (IMIS 12), 2012, pp. 922–926. Article ID 1231430, 2017. 106. Yang, G. et al., A health-IoT platform based on the integration of 88. Hoque, M., and Picard, R. W., Acted vs. Natural frustration and intelligent packaging, unobtrusive bio-sensor, and intelligent med- delight: Many people smile in natural frustration. In Proc. IEEE icine box. IEEE Transactions on Industrial Informatics 10(4): Int. Conf. Autom. Face Gesture Recognit. Workshops (FG), 2011, 2180–2191, 2014. pp. 354–359. 107. Catarinucci, L. et al., An IoT-aware architecture for smart 89. Ray, P. P., Communicating Through Visible Light: Internet of healthcare systems. IEEE Internet of Things 1(2):515–526, 2015. Things Perspective”, Current Science Citation Indexence, Indian 108. Amendola, S. et al., RFID technology for IoT-based patiental Academy of Science, Vol. 111 (12), pp. 1903–1905, 2016. healthcare in smart spaces. IEEE Internet of Things 1(2):144– 90. Hoque, M., and Picard, R. W., Acted vs. Natural frustration and 152, 2014. delight: Many people smile in natural frustration. In Proc. IEEE 109. Barnaghi, P., and Sheth, A., On searching the Internet of Things: Int. Conf. Autom. Face Gesture Recognit. Workshops (FG), 2011, Requirements and challenges. IEEE Intell. Syst. 31(6):71–75, pp. 354–359. 2016. 91. G. Trovato et al., “A novel culture-dependent gesture selection 110. Barnaghi, P., Sheth, A., and Henson, C., From data to actionable system for a humanoid robot performing greeting interaction. In knowledge: Big data challenges in the web of things [guest edi- ’ – Proc. Int. Conf. Social Robot., 2014, pp. 340–349. tors introduction]. IEEE Intell. Syst. 28(6):6 11, 2013. 92. B. Roark, M. Mitchell, and K. Hollingshead, Syntactic complexity 111. Ray, P. P., Creating Values out of Internet of Things: An Industrial measures for detecting mild cognitive impairment. In Proc. Perspective. Journal of Computer Networks and Communications, Workshop Biolo., Transl., Clin. Lang. Process. (BioNLP), 2007, Hindawi, 2016. https://doi.org/10.1155/2016/1579460. pp. 1–8. 112. Shi, J., Zheng, X., Li, Y., Zhang, Q., and Ying, S., Multimodal 93. Hess, U., Adams, R. B., Simard, A., Stevenson, M. T., and Kleck, neuroimaging feature learning with multimodal stacked deep polynomial networks for diagnosis of alzheimer’s disease. IEEE R. E., Smiling and sad wrinkles: Age-related changes in the face – and the perception of emotions and intentions. J. Exp. Soc. Journal of Biomedical and Health Informatics 22(1):173 183, Psychol. 48(6):1377–1380, 2012. 2018. 113. Al-Qazzaz, N. K., Ali, S. H. B. M., Ahmad, S. A., Chellappan, K., 94. Saffran, E. M., Berndt, R. S., and Schwartz, M. F., The quantita- Islam, M. S., and Escudero, J., Role of eeg as biomarker in the tive analysis of agrammatic production: Procedure and data. Brain early detection and classification of dementia. Sci. World J. 2014: Lang. 37(3):440–479, 1989. 906038, 2014. 95. Vermeulen, J., Bastiaanse, R., and Van Wageningen, B., 114. Ray, P. P.,A Survey on Internet of Things Architectures. Journal of Spontaneous speech in aphasia: A correlational study. Brain King Saud University - Computer and Information Sciences, Lang. 36(2):252–274, 1989. Elsevier, 2016. https://doi.org/10.1016/j.jksuci.2016.10.003. 96. Chao, C., and Thomaz, A. L., Turn taking for human–robot inter- 115. Morabito, F. C., Campolo, M., Ieracitano, C., Ebadi, J. M., action. In: Proc. AAAI Fall Symp. Dialog Robots, 2010, pp. 132– Bonanno, L., Bramanti, A., Desalvo, S., Mammone, N., and 134. Bramanti, P., Deep convolutional neural networks for classifica- 97. Tanaka, H. et al., Detecting Dementia Through Interactive tion of mild cognitive impaired and alzheimer’s disease patients Computer Avatars. IEEE Journal of Translational Engineering in from scalp EEG recordings. In 2016 IEEE 2nd International – Health and Medicine 5:1 11, 2017 Art no. 2200111. Forum on Research and Technologies for Society and Industry 98. Solis, M., Digital Tracking of Cognitive Decline: Researchers Are Leveraging a better tomorrow (RTSI), 2016, pp. 1–6. Co-opting Computers in their Efforts to Detect Early Signs of 116. R. Alam, M. Anderson, A. Bankole, and J. Lach, Inferring phys- – Dementia. IEEE Pulse 8(2):16 20, 2017. ical agitation in dementia using smartwatch and sequential behav- 99. Ray, P. P., An IR Sensor Based Smart System to Approximate ior models. In 2018 IEEE EMBS International Conference on Body Core Temperature. Journal of Medical Systems, Springer Biomedical Health Informatics (BHI), 2018, pp. 170–173. 41:123, 2017. 117. Ray, P. P., and Thapa, N., “A Systematic Review on Real-time 100. Ray, P. P., Agarwal, S., “Bluetooth 5 and Internet of Things: Automated Measurement of IV Fluid Level: Status and Potential and Architecture”, IEEE International Conference on Challenges”,Measurement.Elsevier129:343–348, 2018. Signal Processing, Communication & Embedded Systems 118. Ray, P. P., Minimizing Dependency On Internetwork: Is Dew (SCOPES), Paralakhemundi, pp. 1461–1465, 2016. https://doi. Computing a Solution? Transactions on Emerging org/10.1109/SCOPES.2016.7955682 Telecommunications Technologies, Wiley-Blackwell, 101. Ray, P. P., An Internet of Things based Approach to Thermal (Accepted), 30(1), 2019. Comfort Measurement and Monitoring. IEEE International 119. Mendez, M. F., Early-onset alzheimer disease. Neurol. Clin. 35: Conference on Advances in Computing and Communications 263–281, 2017. (ICACCS), pp. 1–7, 2015, 10.1109/ICACCS.2016.7586398, 120. Agosta, F., Caso, F., and Filippi, M., Dementia and neuroimaging. 2016. J. Neurol. 260(2):685–691, 2013. 102. Ray, P. P., Towards an Internet of Things based Architectural 121. Jellinger, K. A., Dementia with lewy bodies and parkinson’s Framework for Defence. IEEE International Conference on diseasedementia: current concepts and controversies. J. Neural Control, Instrumentation, Communication and Computational Transm., (Vienna, Austria: 1996) 125(4):615–650, 2018. Technologies (ICCICCT), Kumaracoil, pp. 411–416. 10.1109/ 122. Ghosh, S., and Lippa, C. F., Clinical subtypes of frontotemporal ICCICCT.2015.7475314, 2015. dementia. Am. J. Alzheimers Dis. Other Dement. 30(7):653–661, 103. Enshaeifar, S. et al., The Internet of Things for Dementia Care. 2015. IEEE Internet Comput. 22(1):8–17, 2018. 123. Ray, P. P., A Generic Internet of Things Architecture for Smart 104. Demiris, G. et al., Technologies for an aging society: a systematic Sports. IEEE International Conference on Control, review of ‘smart home’ applications. Yearbook of Medical Instrumentation, Communication and Computational Informatics 3:33–40, 2008. Technologies x J Med Syst (2019) 43: 287 Page 19 of 21 287

124. Mills, S., Cain, J., Purandare, N., and Jackson, A., Biomarkers of for anatomical brain mri in dementia. Neuroinformatics 14(3): cerebrovascular disease in dementia, The British Journal of 279–296, 2016. Radiology. 80(2):S128–S145, 2007. 139. Rabeh, A. B., Benzarti, F., and Amiri, H., Diagnosis of Alzheimer 125. Liu, M., Zhang, D., and Shen, D., Relationship induced multi- diseases in early step using svm (support vector machine). In 2016 template learning for diagnosis of alzheimer’sdiseaseandmild 13th International Conference on Computer Graphics, Imaging cognitive impairment. IEEE Trans. Med. Imaging 35(6):1463– and Visualization (CGiV), 2016, pp. 364–367. 1474, 2016. 140. Tong, T., Wolz, R., Gao, Q., Guerrero, R., Hajnal, J. V., and 126. Kurth, J., Sakretz, M., Teipel, S., and Krause, B. J., Molecular Rueckert, D., Multiple instance learning for classification of de- imaging of dementia. Geriatric Mental Health Care 1(3):56–62, mentia in brain mri. Med. Image Anal. 18(5):808–818, 2014. 2013. 141. Boser, B. E., Guyon, I. M., and Vapnik, V.N., A training algorithm 127. Valkanova, V., and Ebmeier, K. P., Neuroimaging in dementia. for optimal margin classifiers. Proceedings of Annual Acm Maturitas 79(2):202–208, 2014. Workshop on Computational Learning Theory 5:144–152, 1992. 128. Niessen, W. J., Mr brain image analysis in dementia: From quan- 142. Ray, P. P., Internet of Things for Sports (IoTSport): An architec- titative imaging biomarkers to ageing brain models and imaging tural framework for sports and recreational activity. In Proceeding genetics. Med. Image Anal. 33:107–113, 2016. of IEEE International Conference on Electrical, Electronics, 129. Goujon, A., Mejdoubi, M., Purcell, Y., Banydeen, R., Colombani, Signals, Communication and Optimization (EESCO), Vizag, pp- S., and Arrigo, A., Can mri water apparent diffusion coefficient 79-83, https://doi.org/10.1109/EESCO.2015.7253963, 2015. (adc) value discriminate between idiopathic normal pressure hy- 143. Alkabawi, E. M., Hilal, A. R., and Basir O. A., Feature abstraction drocephalus, alzheimer’s disease and subcortical vascular demen- for early detection of multi-type of dementia with sparse auto- tia? Journal of Neuroradiology. Journal de Neuroradiologie 45(1): encoder. In 2017 IEEE International Conference on Systems, 15–22, 2018. Man, and Cybernetics (SMC), 2017, pp. 3471–3476. 130. Joki, H., Higashiyama, Y., Nakae, Y., Kugimoto, C., Doi, H., 144. Z. Wang, P. J. Markiewicz, G. Platsch, J. Kornhuber, T. Kuwert, Kimura, K., Kishida, H., Ueda, N., Nakano, T., Takahashi, T., and D. Merhof, Comparison of methods for classification of Koyano, S., Takeuchi, H., and Tanaka, F., White matter alzheimer’s disease, frontotemporal dementia and asymptomatic hyperintensities on mri in dementia with lewy bodies, parkinson’s controls. In 2013 IEEE Nuclear Science Symposium and Medical disease with dementia, and alzheimer’s disease. J. Neurol. Sci. Imaging Conference (2013 NSS/MIC), 2013, pp. 1–6. 385:99–104, 2018. 145. Zheng, W., Yao, Z., Hu, B., Gao, X., Cai, H., and Moore, P.,Novel 131. Ray, P. P., Internet of Things based Smart Measurement and cortical thickness pattern for accurate detection of alzheimer’sdis- Monitoring of Wood Equilibrium Moisture Content. IEEE ease. Journal of Alzheimer’s Disease: JAD 48(4):995–1008, 2015. International Conference on Smart Structures and Systems 146. Ray, P. P., Home Health Hub Internet of Things (H3IoT): An (ICSSS), https://doi.org/10.1109/SMARTSENS.2015.7873612, Architectural Framework for Monitoring Health of Elderly 2015. People. In Proceeding of IEEE ICSEMR, ISBN: 132. Kim, J., and Lee, B., Automated discrimination of dementia spec- 9789380222745, Chennai, pp. 1–4, https://doi.org/10.1109/ trum disorders using extreme learning machine and structural t1 ICSEMR.2014.7043542, 2014. mri features. In 2017 39th Annual International Conference of the 147. Litjens, G., Kooi, T., Bejnordi, B. E., Setio, A. A. A., Ciompi, F., IEEE Engineering in Medicine and Biology Society (EMBC), Ghafoorian, M., van der Laak, J. A. W. M., van Ginneken, B., and 2017, pp. 1990–1993. Sánchez, C. I., A survey on deep learning in medical image anal- 133. Zeighami, Y., Fereshtehnejad, S.-M., Dadar, M., Collins, D. L., ysis. Med. Image Anal. 42:60–88, 2017. Postuma, R. B., Mišić, B., and Dagher, A., A clinical-anatomical 148. Ray, P. P., Internet of Things based Physical Activity Monitoring signature of parkinson’s disease identified with partial least (PAMIoT): An Architectural Framework to Monitor Human squares and magnetic resonance imaging. NeuroImage, 2017. Physical Activity. Kolkata: Proceeding of IEEE CALCON, 134. Ramírez, J., Górriz, J., Ortiz, A., Martínez-Murcia, F., Segovia, F., 2014, 32–34. Salas-Gonzalez, D., Castillo-Barnes, D., Illán, I., Puntonet, C., and 149. Alkabawi, E. M., Hilal, A. R., and Basir, O. A., Computer-aided Alzheimer’s Disease Neuroimaging Initiative, Ensemble of ran- classification of multi-types of dementia via convolutional neural dom forests one vs. rest classifiers for mci and ad prediction using networks. In 2017 IEEE International Symposium on Medical anova cortical and subcortical feature selection and partial least Measurements and Applications (MeMeA), 2017, pp. 45–50. squares. J. Neurosci. Methods 302:47–57, 2018. 150. Akhila, J. A., Markose, C., and Aneesh, R. P., Feature extraction 135. Canu, E., Agosta, F., Mandic-Stojmenovic, G., Stojkovic, T., and classification of dementia with neural network. In 2017 Stefanova, E., Inuggi, A., Imperiale, F., Copetti, M., Kostic, V. International Conference on Intelligent Computing, S., and Filippi, M., Multiparametric mri to distinguish early Instrumentation and Control Technologies (ICICICT), 2017, pp. onset alzheimer’s disease andbehavioural variant of 1446–1450. frontotemporal dementia. NeuroImage: Clinical 15:428–438, 151. Dolph, C. V., Alam, M., Shboul, Z., Samad, M. D., and 2017. Iftekharuddin, K. M., Deep learning of texture and structural fea- 136. Ray, P. P., A Generic Internet of Things Architecture for Smart tures for multiclass alzheimer’s disease classification. In 2017 Sports. IEEE International Conference on Control, International Joint Conference on Neural Networks (IJCNN), Instrumentation, Communication and Computational 2017, pp. 2259–2266. Technologies (ICCICCT), Kumaracoil, pp. 405–410, https://doi. 152. Faturrahman, M., Wasito, I., Hanifah, N., and Mufidah, R., org/10.1109/ICCICCT.2015.7475313, 2015. Structural mri classification for alzheimer’s disease detection 137. Harper, L., Fumagalli, G. G., Barkhof, F., Scheltens, P.,O’Brien, J. using deep belief network. in International Conference on T., Bouwman, F., Burton, E. J., Rohrer, J. D., Fox, N. C., Ridgway, Information & Communication Technology and System, 2017, G. R., and Schott, J. M., Mri visual rating scales in the diagnosis of pp. 37–42. dementia: evaluation in 184 post-mortem confirmed cases. Brain 153. Ray, P. P., Nano Computer Design Based on Intra Body Nanoscale 139(4):1211–1225, 2016. Neuro-Spike Communication: a Nanonetwork Paradigm. ISBN 138. Tohka, J., Moradi, E., Huttunen, H., and Initiative, A.’s. D. N., 978–1–4673-4698-6. In proceeding of IEEE International Comparison of feature selection techniques in machine learning Conference on Communications, Devices and Intelligent 287 Page 20 of 21 J Med Syst (2019) 43: 287

Systems (CODIS), JU, Kolkata, pp. 314–317. https://doi.org/10. behavior models,” in 2018 IEEE EMBS International 1109/CODIS.2012.6422199,2012. Conference on Biomedical Health Informatics (BHI), 2018, pp. 154. Suk, H.-I., Lee, S.-W., Shen, D., and Alzheimer’s Disease 170–173. Neuroimaging Initiative, Deep ensemble learning of sparse re- 168. 2018 alzheimer’s disease facts and figures. Alzheimer’s& gression models for brain disease diagnosis. Med. Image Anal. Dementia. 14(3): 367–429, 2018. 37:101–113, 2017. 169. World health organization, http://www.who.int/news-room/facts- 155. Islam, J., and Zhang, Y., A novel deep learning based multi-class inpictures/detail/dementia, Accessed: 2018-05-25. classification method for alzheimer’s disease detection using brain 170. Mendez, M. F., Early-onset alzheimer disease. Neurol. Clin. 35: MRI data. In International Conference on Brain Informatics, 2017, 263–281, 2017. pp. 213–222. 171. Agosta, F., Caso, F., and Filippi, M., Dementia and neuroimaging. 156. Liu, J., Shang, S., Zheng, K., and Wen, J.-R., Multi-view ensemble J. Neurol. 260(2):685–691, 2013. learning for dementia diagnosis from neuroimaging: An artificial 172. Jellinger, K. A., Dementia with lewy bodies and parkinson’s neural network approach. Neurocomputing 195:112–116, 2016. diseasedementia: current concepts and controversies. J. Neural – 157. Ithapu, V. K., Singh, V., Okonkwo, O. C., Chappell, R. J., Transm. (Vienna, Austria: 1996) 125(4):615 650, 2018. Dowling, N. M., and Johnson, S. C., Imaging-based enrichment 173. Ghosh, S., and Lippa, C. F., Clinical subtypes of frontotemporal – criteria using deep learning algorithms for efficient clinical trials dementia. Am. J. Alzheimers Dis. Other Dement. 30(7):653 661, in mild cognitive impairment. Alzheimers Dement. 11(12):1489– 2015. ’ 1499, 2015. 174. Iwatsubo, T., Alzheimer s disease neuroimaging initiative 158. Choi, H., Deep learning in nuclear medicine and molecular imag- (ADNI). Japanese Journal of Clinical Medicine 69(Suppl 8): – ing: Current perspectives and future directions. Nucl. Med. Mol. 570 574, 2011. Imaging 52(2):109–118, 2018. 175. Mills, S., Cain, J., Purandare, N., and Jackson, A., Biomarkers of cerebrovascular disease in dementia. Br. J. Radiol. 80(2):S128– 159. Lu, D., Popuri, K., Ding, G. W., Balachandar, R., Beg, M. F., and S145, 2007. Alzheimer’s Disease Neuroimaging Initiative, Multimodal and 176. Liu, M., Zhang, D., and Shen, D., Relationship induced multi- multiscale deep neural networks for the early diagnosis of template learning for diagnosis of alzheimer’sdiseaseandmild alzheimer’s disease using structural MR and FDG-pet images. cognitive impairment. IEEE Trans. Med. Imaging 35(6):1463– Sci. Rep. 8(1):5697, 2018. 1474, 2016. 160. Lu, D., Popuri, K., Ding, G. W., Balachandar, R., Beg, M. F., and 177. Kurth, J., Sakretz, M., Teipel, S., and Krause, B. J., Molecular ADNI, Multiscale deep neural network based analysis of fdg-pet imaging of dementia. Geriatric Mental Health Care 1(3):56–62, images for the early diagnosis of alzheimer’s disease. Med. Image 2013. Anal. 46:26–34, 2018. 178. Valkanova, V., and Ebmeier, K. P., Neuroimaging in dementia. 161. Ju, R., Hu, C., Zhou, P., and Li, Q., “Early diagnosis of alzheimer’s Maturitas 79(2):202–208, 2014. disease based on resting-state brain networks and deep learning,” 179. Niessen, W. J., Mr brain image analysis in dementia: From quan- IEEE/ACM Transactions on Computational Biology and titative imaging biomarkers to ageing brain models and imaging Bioinformatics, pp. 1–1, 2017. genetics. Med. Image Anal. 33:107–113, 2016. 162. Anazodo, U. C., Finger, E., Kwan, B. Y. M., Pavlosky, W., 180. Henriksen, O. M., Marner, L., and Law, I., Clinical pet/mr imaging Warrington, J. C., Günther, M., Prato, F. S., Thiessen, J. D., and in dementia and neuro-oncology. PET Clinics 11(4):441–452, Lawrence, K. S. S., Using simultaneous pet/mri to compare the 2016. accuracy of diagnosing frontotemporal dementia by arterial spin 181. Kato,T.,Inui,Y.,Nakamura,A.,andIto,K.,Brain – labelling mri and fdg-pet. NeuroImage. Clinical 17:405 414, fluorodeoxyglucose (fdg) pet in dementia. Ageing Res. Rev. 30: 2018. 73–84, 2016. 163. Shi, J., Zheng, X., Li, Y., Zhang, Q., and Ying, S., Multimodal 182. Valkanova, V., and Ebmeier, K. P., Neuroimaging in dementia. neuroimaging feature learning with multimodal stacked deep Maturitas 79(2):202–208, 2014. ’ polynomial networks for diagnosis of alzheimer s disease. IEEE 183. Ray, P. P., An Introduction to Dew Computing: Definition, – Journal of Biomedical and Health Informatics 22(1):173 183, Concept and Implications. IEEE Access 6:723–737, 2017. 2018. 184. Arab, A., Wojna-Pelczar, A., Khairnar, A., Szabó, N., and Ruda- 164. Al-Qazzaz, N. K., Ali, S. H. B. M., Ahmad, S. A., Chellappan, K., Kucerova, J., Principles of diffusion kurtosis imaging and its role Islam, M. S., and Escudero, J., Role of eeg as biomarker in the in early diagnosis of neurodegenerative disorders. Brain Res. Bull. early detection and classification of dementia. 139:91–98, 2018. TheScientificWorldJournal 2014:906038, 2014. 185. Jack, C. R., Bernstein, M. A., Fox, N. C., Thompson, P., 165. Morabito, F. C., Campolo, M., Ieracitano, C., Ebadi, J. M., Alexander, G., Harvey, D., Borowski, B., Britson, P. J., Bonanno, L., Bramanti, A., Desalvo, S., Mammone, N., and Whitwell, J. L., and Ward, C., The alzheimer’s disease neuroim- Bramanti, P., Deep convolutional neural networks for classifica- aging initiative (adni). J. Magn. Reson. Imaging 9(1):101, 2015. tion of mild cognitive impaired and alzheimer’sdiseasepatients 186. Marcus, D. S., Wang, T. H., Parker, J., Csernansky, J. G., Morris, J. from scalp eeg recordings. In 2016 IEEE 2nd International Forum C., and Buckner, R. L., Open access series of imaging studies on Research and Technologies for Society and Industry (oasis): Crosssectional mri data in young, middle aged, Leveraging a better tomorrow (RTSI), 2016, pp. 1–6. nondemented, and demented older adults. J. Cogn. Neurosci. 166. Alam, R., Dugan, J., Homdee, N., Gandhi, N., Ghaemmaghami, 19(9):1498–1507, 2007. B.,Meda,H.,Bankole,A.,Anderson,M.,Gong,J.,Smith- 187. Villemagne, V. L., Aibl (australian imaging, biomarkers and life- Jackson, T., and Lach, J., Besi: Reliable and heterogeneous sens- style) flagship study of aging. Alzheimers Dement. 5(4):P1–P2, ing and intervention for in-home health applications. In 2017 2009. IEEE/ACM International Conference on Connected Health: 188. Caddementia: A standardized evaluation framework for computer- Applications, Systems and Engineering Technologies (CHASE), aided diagnosis of dementia based on structural MRI. https:// 2017, pp. 147–156. caddementia.grand-challenge.org/Home/, Accessed: 2018-05-26. 167. R. Alam, M. Anderson, A. Bankole, and J. Lach, “Inferring phys- 189. Bhuvaneshwari, B and Kavitha A., Investigations on the brain ical agitation in dementia using smartwatch and sequential connectivity patterns in progression of alzheimer’s disease using J Med Syst (2019) 43: 287 Page 21 of 21 287

functional mr imaging and graph theoretical measures. in 2017 disease with dementia, and alzheimer’s disease. J. Neurol. Sci. IEEE 16th International Conference on Cognitive Informatics & 385:99–104, 2018. Cognitive Computing (ICCI*CC), , 2017, pp. 151–160. 199. Ray, P. P., Dash, D., and De, D., Analysis and Monitoring of IoT 190. Ray, P. P.,Dash, D., and De, D., A Systematic Review of Wearable Assisted Human Physiological Galvanic Skin Response Factor for Systems for Cancer Detection: Current State and Challenges. J. Smart E-Healthcare. Sens. Rev., Emerald Publishing, 2019. Med. Syst. 41:180, 2017. https://doi.org/10.1108/SR-07-2018-0181. 191. Sandhya, C., Vidhusha, S., Gayathri, R. S., Sneha Priya, B., and 200. Islam, A. J., Ahammad, K. T., Barua, B., Alam, S. S., and Biswas, Kavitha, A., Assessment of functional connectivity in autistic A., A new approach of brain mri analysis for identifying brain FMRI. In 2015 IEEE 14th International Conference on creutzfeldtjakob disease (cjd). In 2017 3rd International Cognitive Informatics Cognitive Computing (ICCI*CC), 2015, Conference on Electrical Information and Communication pp. 169–174. Technology (EICT), 2017, pp. 1–5. 192. Ray, P. P., Internet of Things for Smart Agriculture: Technologies, 201. Majumder, P., Ray, P. P., Ghosh, S., and Dey, S. K., Potential Practices and Future Road Map. Journal of Ambient Intelligence Effect of Tobacco Consumption Through Smoking and Chewing and Smart Environments, IOS Press 9:395–420, 2017. Tobacco on IL1beta Protein Expression in Chronic Periodontitis 193. Anandh, K. R., Sujatha, C. M., and Ramakrishnan, S., Analysis of Patients: In Silico Molecular Docking Study. IEEE/ACM ventricles in alzheimer mr images using coherence enhancing dif- Transactions on Computational Biology and Bioinformatics, fusion filter and level set method. In 2014 International 2019. https://doi.org/10.1109/TCBB.2019.2894737. Conference on Informatics, Electronics Vision (ICIEV), 2014, 202. Huang, W., Zeng, S., Li, J., and Chen, G., A new image-based pp. 1–4. immersive tool for dementia diagnosis using pairwise ranking and – 194. Cajanus, A., Hall, A., Koikkalainen, J., Solje, E., Tolonen, A., learning. Multimed. Tools Appl. 75(9):5359 5376, 2016. Urhemaa, T., Liu, Y., Haanpää, R. M., Hartikainen, P., 203. Ray, P. P., Dash, D., and De, D., Internet of Things-based Real- Helisalmi, S., Korhonen, V., Rueckert, D., Hasselbalch, S., Time Model Study on e-Healthcare: Device, Message Service and – Waldemar, G., Mecocci, P., Vanninen, R., van Gils, M., Dew Computing. Comput. Netw., Elsevier 149:226 239, 2019. Soininen, H., Lötjöonen, J., and Remes, A. M., Automatic mri 204. Steketee, R. M. E., Bron, E. E., Meijboom, R., Houston, G. C., quantifying methods in behavioral-variant frontotemporal demen- Klein, S., Mutsaerts, H. J. M. M., Mendez Orellana, C. P., de Jong, tia diagnosis. Dementia and Geriatric Cognitive Disorders Extra F. J., van Swieten, J. C., van der Lugt, A., and Smits, M., Early- ’ 8(1):51–59, 2018. stage differentiation between presenile alzheimer s disease and frontotemporal dementia using arterial spin labeling MRI. Eur. 195. Ray, P. P., Internet of Things based Approximation of Sun Radiol. 26(1):244–253, 2016. Radiative-Evapotranspiration (ET0) Models. Journal of 205. Wierenga, C. E., Hays, C. C., and Zlatar, Z. Z., Cerebral blood Agrometeorology 20(2):171–173, 2018. flow measured by arterial spin labeling mri as a preclinical marker 196. Ray, P. P., Continuous Glucose Monitoring: A Review of Sensor of alzheimer’s disease. Journal of Alzheimer’sDisease:JAD Systems and Prospects. Sens. Rev. 38(4):420–437. 42(Suppl4):S411 –S419, 2014. 197. Goujon, A., Mejdoubi, M., Purcell, Y., Banydeen, R., Colombani, 206. Ray, P. P., Dash, D., and De, D., Approximation of Fruit Ripening S., and Arrigo, A., Can mri water apparent diffusion coefficient Quality Index for IoT based Assistive e-Healthcare. Microsystem (adc) value discriminate between idiopathic normal pressure hy- Technologies, Springer, 2018. drocephalus, alzheimer’s disease and subcortical vascular dementia? Journal of Neuroradiology. Journal de Neuroradiologie 45(1): 15–22, 2018. 198. Joki, H., Higashiyama, Y., Nakae, Y., Kugimoto, C., Doi, H., Publisher’sNote Springer Nature remains neutral with regard to jurisdic- Kimura, K., Kishida, H., Ueda, N., Nakano, T., Takahashi, T., tional claims in published maps and institutional affiliations. Koyano, S., Takeuchi, H., and Tanaka, F., White matter hyperintensities on mri in dementia with lewy bodies, parkinson’s