Passive Monitoring Techniques Research Articles

Dietary Assessment with Multimodal ChatGPT: A Systematic Analysis.

Conventional approaches to dietary assessment are primarily grounded in self-reporting methods or structured interviews conducted under the supervision of dietitians. These methods, however, are often subjective, potentially inaccurate, and time-intensive. Although artificial intelligence (AI)-based solutions have been devised to automate the dietary assessment process, prior AI methodologies tackle dietary assessment in a fragmented landscape (e.g., merely recognizing food types or estimating portion size), and encounter challenges in their ability to generalize across a diverse range of food categories, dietary behaviors, and cultural contexts. Recently, the emergence of multimodal foundation models, such as GPT-4V, has exhibited transformative potential across a wide range of tasks (e.g., scene understanding and image captioning) in various research domains. These models have demonstrated remarkable generalist intelligence and accuracy, owing to their large-scale pre-training on broad datasets and substantially scaled model size. In this study, we explore the application of GPT-4V powering multimodal ChatGPT for dietary assessment, along with prompt engineering and passive monitoring techniques. We evaluated the proposed pipeline using a self-collected, semi free-living dietary intake dataset comprising 16 real-life eating episodes, captured through wearable cameras. Our findings reveal that GPT-4V excels in food detection under challenging conditions without any fine-tuning or adaptation using food-specific datasets. By guiding the model with specific language prompts (e.g., African cuisine), it shifts from recognizing common staples like rice and bread to accurately identifying regional dishes like banku and ugali. Another GPT-4V's standout feature is its contextual awareness. GPT-4V can leverage surrounding objects as scale references to deduce the portion sizes of food items, further facilitating the process of dietary assessment.

Comprehensive assessment and monitoring of bond deterioration in GFRP-reinforced concrete beams using guided wave and acoustic emission techniques

The debonding between glass fiber-reinforced polymer (GFRP) rebars and concrete significantly reduces the bond strength and jeopardizes the safety and durability of GFRP-reinforced concrete (GFRP-RC) structures. As a result, it is imperative to monitor the structural integrity of infrastructures being built using these composite materials in structural elements during their service lives. Previous studies have primarily focused on artificially induced debonding scenarios for guided wave (GW)-based monitoring, which has resulted in an inability to capture debonding events that arise naturally due to external loads. Furthermore, flexural bond test investigations simulating real loading scenarios in GFRP-RC structures have not been linked with acoustic emission (AE) and GW monitoring. The present study addresses these research gaps by comprehensively examining the flexural bond integrity in GFRP-RC structural elements subjected to realistic loading conditions using a hybrid approach combining GW and AE as active and passive structural health monitoring techniques, respectively. A series of bond tests, conforming to RILEM specifications, have been performed to investigate various parameters (including rebar surface characteristics, embedment length, and confinement conditions) that affect the flexural bond degradation response of GFRP-RC members. Damage indices based on GW and AE signal parameters have been developed to evaluate the characteristics of confined and unconfined specimens distinctly. It has been observed from the test results that the L (0,3) GW mode is more sensitive to debonding than other modes and that a threshold of 25% has been established for the GW damage index to denote significant bond damage. Furthermore, AE damage indices have provided valuable insights into concrete damage, including splitting and damage at the rebar-concrete interface. Thus, a thorough understanding of bond deterioration in GFRP-RC structures has been developed by combining GW and AE health monitoring techniques, with implications for enhancing structural safety and durability.

Investigation of duty cycles for measuring activity in passive acoustic bat monitoring

Echolocating bats are important bioindicators that can be monitored effectively using passive acoustic monitoring (PAM) techniques. In PAM, duty-cycle-based temporal subsampling is often used to collect data at ON/OFF intervals to circumvent the limitations of recorder battery and storage capacity to enable long-term monitoring. However, potential bias introduced by temporal subsampling has not been systematically investigated for bat monitoring. Here, we use continuous audio recordings from the Union Bay Natural Area in Seattle in Summer 2022 to simulate the effects of temporal subsampling using different duty cycle parameters. We detected bat calls automatically using a deep learning model [Aodha et al., 2022, BioRxiv] and calculated three metrics as proxy for bat activity: number of calls, Activity Index (AI), and Bout Time Percentage (BTP). We found that although the number of calls and AI are more readily computable using the detected calls, BTP is likely a more accurate measure that relies less on the performance of automated call detection methods. In addition, reduced sampling for only a portion of the night (e.g., 4 hrs) was generally inadequate for capturing bat activity. Our results suggest that considering species-specific acoustic characteristics is crucial for reducing sampling bias for PAM of bats.

Smart surveillance using IoT: a review

In today’s modern society, video surveillance is a growing trend and it can revolutionize many aspects of technology, especially in future smart cities that will transform traditional surveillance systems into intelligent and interconnected networks. It may be difficult for even well-trained employees to process and respond immediately to monitored data. Moreover, IoT-enabled surveillance systems overcome the challenges and flaws of conventional passive monitoring techniques by offering real-time surveillance and automated notifications for suspicious activity, intrusions, or anomalies. Therefore, the objective of this article is to provide an Internet of Things (IoT)-based smart surveillance system that may successfully connect the ecosystem, resulting in enhanced monitoring for smart city services. The primary goal of this study is to provide a comprehensive review of several IoT-based surveillance techniques used in various smart city applications. The categorization of tasks for each section is as follows: to present the historical context overview; to examine the significance of IoT and its application in smart cities; to present a standardized architecture for IoT-based smart city surveillance; to highlight an authoritative and thorough review of current IoT-based smart surveillance systems; and to identify various research issues. The methods used are: statistical graph or chart approach, schematic, and timeline diagram. Conclusions. This article outlines numerous research challenges for future video surveillance that may be addressed by researchers. In summary, this comprehensive review provides a valuable and streamlined resource for future researchers exploring smart city surveillance through the IoT.

Investigating VOCs Speciation Characteristics at the Fenceline of Synthetic Rubber Manufacturing Industries via Active and Passive Monitoring Techniques

Investigating VOCs Speciation Characteristics at the Fenceline of Synthetic Rubber Manufacturing Industries via Active and Passive Monitoring Techniques

Laboratory investigation of hydraulic fracturing in granitic rocks using active and passive seismic monitoring

SUMMARY Knowledge of the fracturing processes can be important for the optimization of pressurized fluid injection operations in the deep underground rock mass. Active and passive seismic monitoring techniques have been used in the field for tracking or mapping the propagating hydraulic fracture. Although both these monitoring techniques provide valuable information about the generated fracture network, it is difficult for either technique to comprehensibly identify the different processes associated with hydraulic fracturing. The combined active and passive monitoring has the potential for better characterization of the complex hydraulic fracturing phenomena. In this study, laboratory hydraulic fracturing experiments with combined active and passive seismic monitoring were conducted on true triaxially loaded Barre granite cubes with different fluid injection rates. The seismic inelastic fracturing was detected by 16 passive acoustic emission sensors, where 3678 and 2370 seismic source events were detected for the high and low injection rate experiments, respectively. For active monitoring, strong variations in the attributes of signals were observed which were transmitted through four source–receiver pairs, placed both perpendicular and parallel to the generated hydraulic fracture. Positive velocity changes were observed for active sensor pairs with ray paths passing through the generated hydraulic fracture indicating fluid permeation, whereas isolated dry deformation was characterized by a slight but permanent velocity decrease. Compared to velocity, the energy of the active signals was 1–2 orders of magnitude more sensitive to different hydraulic fracturing processes. However, the sensitivity and signatures of the active signal attributes were found to be dependent on the frequency range and direction of ray path with respect to the location of the generated fracture network. Using the coupled evaluation of the active and passive signals we were able to systematically identify various hydraulic fracturing processes including: (1) aseismic deformation, (2) fracture initiation and fluid permeation, (3) pressure build-up, (4) fracture propagation and (5) pressure release and leak-off. The results of this study showed that combining the respective advantages of active and passive seismic techniques and using both of them to monitor the failure processes can facilitate a more comprehensive understanding and better control of the hydraulic stimulations in subsurface operations.

Anthropogenic activity, hydrological regime, and light level jointly influence temporal patterns in biosonar activity of the Yangtze finless porpoise at the junction of the Yangtze River and Poyang Lake, China.

Under increasing anthropogenic pressure, species with a previously contiguous distribution across their ranges have been reduced to small fragmented populations. The critically endangered Yangtze finless porpoise ( Neophocaena asiaeorientalis asiaeorientalis), once commonly observed in the Yangtze River-Poyang Lake junction, is now rarely seen in the river-lake corridor. In this study, static passive acoustic monitoring techniques were used to detect the biosonar activities of the Yangtze finless porpoise in this unique corridor. Generalized linear models were used to examine the correlation between these activities and anthropogenic impacts from the COVID-19 pandemic lockdown and boat navigation, as well as environmental variables, including hydrological conditions and light levels. Over approximately three consecutive years of monitoring (2020-2022), porpoise biosonar was detected during 93% of logged days, indicating the key role of the corridor for finless porpoise conservation. In addition, porpoise clicks were recorded in 3.80% of minutes, while feeding correlated buzzes were detected in 1.23% of minutes, suggesting the potential existence of localized, small-scale migration. Furthermore, both anthropogenic and environmental variables were significantly correlated with the diel, lunar, monthly, seasonal, and annual variations in porpoise biosonar activities. During the pandemic lockdown period, porpoise sonar detection showed a significant increase. Furthermore, a significant negative correlation was identified between the detection of porpoise click trains and buzzes and boat traffic intensity. In addition to water level and flux, daylight and moonlight exhibited significant correlations with porpoise biosonar activities, with markedly higher detections at night and quarter moon periods. Ensuring the spatiotemporal reduction of anthropogenic activities, implementing vessel speed restrictions (e.g., during porpoise migration and feeding), and maintaining local natural hydrological regimes are critical factors for sustaining porpoise population viability.

Combining Active and Passive Stress Wave Monitoring Techniques: Opportunities for Condition Assessment of Concrete Structures

Condition monitoring of concrete structures has experienced increasing interest over the last decades, especially for bridges. In this presentation, a combined approach based on active and passive stress wave monitoring techniques is proposed and discussed. While from a wave propagation mechanics point of view, acoustic emission (AE) and ultrasonic monitoring are the same, they differ fundamentally in what the cause of the stress wave is. For the former, external stimuli such as overloading or temperature cause the spontaneous release of fracture energy in unknown locations. As a result, AE events cannot be repeated. For the latter, a known source at a known location is employed at a specified time. In contrast, this approach is highly controlled and repeatable. The presenter argues that a combination of these two techniques has the potential to provide a much more comprehensive picture of ongoing fracture processes as well as slowly occurring changes in the structure. An overview and comparison of the techniques is provided, along with observations and results from laboratory experiments and field tests that the presenter has been involved in over the last 15 years.

Seasonal and diel activities of the Yangtze finless porpoise in natural and highly disturbed habitats: Implications for conservation planning of freshwater cetaceans

Abstract Knowledge of fine‐scale seasonal and diel activity patterns of animals provides insights into habitat utilization dynamics and informs conservation action planning against potential human impacts. Such information is particularly important for riverine cetaceans, such as the Yangtze finless porpoise, which is threatened by habitat loss and disturbances by a variety of human activities. Passive acoustic monitoring (PAM) techniques were used to survey the seasonal and diel activities of the Yangtze finless porpoise in two natural habitat patches of the Anqing Yangtze Finless Porpoise Natural Reserve, and in one highly disturbed area adjacent to the reserve. In the two natural habitat patches, significantly higher detection rates of porpoises were observed during autumn and winter. The detection rates during autumn and winter were significantly higher at night than during daytime, whereas during spring and summer the diel difference in porpoise detection rates was not statistically significant. In the highly disturbed habitat, the porpoise detection rates were significantly lower than those in the two natural habitats. Although seasonal differences in porpoise detection rates were not statistically significant, diel detection rates were significantly different in autumn and winter. Information on seasonal patterns of porpoise activities suggested strengthening monitoring and law enforcement during autumn and winter, during the period in which diel activity patterns implied a potential conservation gap during nighttime. Extremely low levels of porpoise activities in port areas implied an ecological impact of coastal engineering on megafauna viability through the alteration of habitat functionality. To provide information on animal occurrences and activities and to minimize the bias of ‘false‐absence’ from visual surveys, the census survey design should consider both visual and PAM techniques. In seasons when occurrences of river cetaceans and habitat use are increasing, heightened conservation monitoring and law enforcement efforts are required, particularly during nighttime.

Additive Manufacturing of Miniaturized Peak Temperature Monitors for In-Pile Applications.

Passive monitoring techniques have been used for peak temperature measurements during irradiation tests by exploiting the melting point of well-characterized materials. Recent efforts to expand the capabilities of such peak temperature detection instrumentation include the development and testing of additively manufactured (AM) melt wires. In an effort to demonstrate and benchmark the performance and reliability of AM melt wires, we conducted a study to compare prototypical standard melt wires to an AM melt wire capsule, composed of printed aluminum, zinc, and tin melt wires. The lowest melting-point material used was Sn, with a melting point of approximately 230 °C, Zn melts at approximately 420 °C, and the high melting-point material was aluminum, with an approximate melting point of 660 °C. Through differential scanning calorimetry and furnace testing we show that the performance of our AM melt wire capsule was consistent with that of the standard melt-wire capsule, highlighting a path towards miniaturized peak-temperature sensors for in-pile sensor applications.

A New Method for Standardizing Inland Fish Community Surveys: Characterizing Habitat Associated with Small-Bodied Fish Species, Abundance, and Size Distributions in a Highly Modified Estuary

Monitoring fish assemblages in estuary environments is often difficult to standardize due to habitat complexity and gear-specific selectivity. This is further compounded by conventional sampling methods which require fish handling that might cause stress or mortality. To ameliorate these issues, we developed the Single-Platform Aquatic Species and Habitat Sampling System (Platform), an integrated and mobile concentrator net and live box prototype, coupled with a range of physical monitoring equipment. The long-term goal of the Platform is to employ non-invasive sampling techniques such as video and eDNA; however, our aim in this study was to test the Platform’s ability to sample across different habitat types and detect differences in fish assemblages. We investigated the utility of the Platform over a short time period (2 mo) in a relatively small and complex embayment within the highly modified San Francisco Estuary, California. We were able to identify clear physical distinctions among estuary/delta habitat types and detect habitat segregation by ecological groups of sampled fishes with the Platform. The Platform also detected discernable ontogenetic shifts (i.e., size differences) within estuary habitats of commonly observed fish species. These initial results demonstrate the Platform’s ability to contrast fish density, size, and species diversity, which sets the stage to advance more passive monitoring techniques, including video and eDNA methods. The Platform has the potential to fill a methodological gap in non-invasive surveying of small-bodied fish across a range of estuarine habitats, warranting further investigation of potential applications.

Estimating Livestock Grazing Activity in Remote Areas Using Passive Acoustic Monitoring

Grazing has long been recognized as an effective means of modifying natural habitats and, by extension, as a wildlife and protected area management tool, in addition to the obvious economic value it has for pastoral communities. A holistic approach to grazing management requires the estimation of grazing timing, frequency, and season length, as well as the overall grazing intensity. However, traditional grazing monitoring methods require frequent field visits, which can be labor intensive and logistically demanding to implement, especially in remote areas. Questionnaire surveys of farmers are also widely used to collect information on grazing parameters, however there can be concerns regarding the reliability of the data collected. To improve the reliability of grazing data collected and decrease the required labor, we tested for the first time whether a novel combination of autonomous recording units and the semi-automated detection algorithms of livestock vocalizations could provide insight on grazing activity at the selected areas of the Greek Rhodope mountain range. Our results confirm the potential of passive acoustic monitoring (PAM) techniques as a cost-efficient method for acquiring high resolution spatiotemporal data on grazing patterns. Additionally, we evaluate the three algorithms that we developed for detecting cattle, sheep/goat, and livestock bell sounds, and make them available to the broader scientific community. We conclude with suggestions on ways that acoustic monitoring can further contribute to managing legal and illegal grazing, and offer a list of priorities for related future research.

Dynamic mode decomposition: A feature extraction technique based hidden Markov model for detection of Mysticetes' vocalisations

The detection and classification of Mysticetes' vocalisations have evoked the attention of researchers over the years because of their relevance to the marine ecosystem. These vocalisations are gathered employing different passive acoustic monitoring techniques. The vocalisation datasets are accumulated over a period; thus, they are large and cannot be easily analysed manually. Consequently, efficient machine learning (ML) tools such as the hidden Markov models (HMMs) are used extensively to automatically detect and classify these huge vocalisation datasets. As with most ML tools, the detection efficiency of the HMMs depend on the adopted feature extraction technique. Feature extraction techniques such as the Mel-scale frequency cepstral coefficient (MFCC), linear predictive coefficient (LPC), and empirical mode decomposition (EMD) have been adopted with the HMMs to detect different Mysticetes' vocalisations. This article introduces the method of dynamic mode decomposition (DMD) as a feature extraction technique that can be adopted with the HMMs for the detection of Mysticetes' vocalisations. The DMD has emerged as a robust tool for analysing the dynamics of non-stationary and non-linear signals. It is a completely data-driven tool that decomposes a signal over a certain period of time into relevant modes. Here, these spatial-temporal modes are reconstructed mathematically to form reliable feature vectors of the decomposed vocalisation signals. The performance of the proposed DMD-HMM detection technique is demonstrated using the acoustic datasets of two different Mysticetes' vocalisations: Humpback whale songs and Inshore Bryde's whale short pulse calls. In both species, the proposed DMD-HMM exhibits superior sensitivity and false discovery rate performances as compared to the MFCC-HMM, LPC-HMM, and EMD-HMM detection methods. Likewise, this proposed DMD-HMM detection method can be extended to other Mysticetes' that produce characteristics sounds.

Beyond Stationarity: Influence of Flow History and Sediment Supply on Coarse Bedload Transport

AbstractUnderstanding how river bedload responds to climate and land use changes and water resource management initiatives is critical in developing sustainable approaches to river management. Passive monitoring techniques permit investigation of interannual dependencies in bedload transport in high resolution, including sediment supply factors. Here, seismic impact plate records are processed using a probabilistic model BedLoad from Impact Plates model to derive a 5‐year bedload data set at 5‐min intervals for the lower River Avon, Devon, UK. For water years that range from very dry to very wet, annual coarse bedload yields are estimated to vary through two orders of magnitude with wide prediction intervals. The most effective discharge occurs consistently at about one‐third of bankfull flow, morphologically at “subbarfull” stage, the result of hysteretic trends and falling limb transport in this non‐threshold channel. A two‐phase sediment rating curve is revealed with a variable supply related “bulge” during in‐bank flows, giving way to a near‐linear trend during overbank flows. The supply related component is predicted well using a sensitivity style metric that combines the cumulative duration of competent flows with the magnitude‐duration product of near‐threshold flows, defining a field‐scale exemplar of “stress history.” Further, the relative proportion of supply related coarse bedload yield relates strongly to the relative wetness of the previous year. High resolution, multiyear data reveal that controls on bedload dynamics are unique to a site's hydrogeoclimatic context and position in the river basin. Passive monitoring holds promise for generating “type sites” of bedload behavior critical for use in improving aquatic biodiversity and the sustainability of river management.

Ecoacoustics: acoustic sensing for biodiversity monitoring at scale

Ecoacoustics: acoustic sensing for biodiversity monitoring at scale

0310 Oculometrics Track Performance on the Psychomotor Vigilance Task During Acute Sleep Deprivation

Abstract Introduction Sleep deprivation and circadian misalignment impairs human sensorimotor performance and reduces vigilant attention, which increases the potential for errors in occupations that require 24-hour operations. The psychomotor vigilance task (PVT) is the gold-standard measure for evaluating the impact of sleepiness on performance, however, it is not practical to administer in many operational environments, because it only provides a snapshot of performance and requires an individual to focus on the task for several minutes, multiple times over a work shift. As a result, passive, continuous monitoring of sleepiness is desirable for operational environments. The goal of the present study was to determine if complex oculomotor behavioral metrics track PVT performance during sleep deprivation. Methods Twelve healthy adults (mean age 24.8 ± 5.4 years; 6F) maintained a fixed schedule with 8.5 hours in bed for two weeks, during which they abstained from caffeine, alcohol, and other medications, followed by a ~24 hours constant routine laboratory stay. Participants completed the PVT and a radial step-ramp ocular tracking task hourly throughout the study. Twelve oculometrics were derived from smooth pursuit and saccadic eye movements collected through video-oculography and were compared to the PVT and Karolinska Sleepiness Scale (KSS) using linear regression and receiver operating characteristic curves. Results Nine oculometrics spanning pursuit, saccade, and directional motion processing performance correlated with the PVT and KSS (p < 0.05), including: (a) pursuit latency; (b) open-loop pursuit acceleration; (c) proportion smooth; (d) steady-state pursuit gain; (e) saccadic amplitude; (f) saccadic dispersion; (g) saccadic rate; (h) direction asymmetry; and (i) direction noise. Conclusion The oculometrics that we examined exhibited a distinct pattern that tracked PVT performance. Future studies should examine whether these metrics can be extracted through passive monitoring techniques. Support None

Acoustic vs. photographic monitoring of gray wolves (Canis lupus): a methodological comparison of two passive monitoring techniques

Remote camera traps are often used in large-mammal research and monitoring programs because they are cost-effective, allow for repeat surveys, and can be deployed for long time periods. Statistical advancements in calculating population densities from camera-trap data have increased the popularity of camera usage in mammal studies. However, drawbacks to camera traps include their limited sampling area and tendency for animals to notice the devices. In contrast, autonomous recording units (ARUs) record the sounds of animals with a much larger sampling area but are dependent on animals producing detectable vocalizations. In this study, we compared estimates of occupancy and detectability between ARUs and remote cameras for gray wolves (Canis lupus Linnaeus, 1758) in northern Alberta, Canada. We found ARUs to be comparable with cameras in their detectability and occupancy of wolves, despite only operating for 3% of the time that cameras were active. However, combining cameras and ARUs resulted in the highest detection probabilities for wolves. These advances in survey technology and statistical methods provide innovative avenues for large-mammal monitoring that, when combined, can be applied to a broad spectrum of conservation and management questions, provided assumptions for these methods are rigorously tested and met.

Fin whale 40‐Hz calling behavior studied with an acoustic tracking array

Fin whale 40‐Hz calling behavior studied with an acoustic tracking array

Passive monitoring techniques to evaluate environmental pesticide exposure: Results from the Infant's Environmental Health study (ISA)

BackgroundPesticides used in agriculture may expose populations living nearby. Costa Rica is a major banana-exporting country, its production depends on extensive pesticide use. ObjectivesTo evaluate environmental pesticide exposure, we measured levels of current-use pesticides in air and dust from 12 schools in Matina County, Costa Rica, with passive sampling methods. MethodsWe selected ten proximal and two non-proximal schools and placed polyurethane foam passive air samplers outdoors at each school, during four consecutive periods. At three of these schools, we also placed an active air sampler during the first 24 h of each sampling period. We collected passive dust samples by placing a glass Petri Dish at the inside of each school. We subsequently performed a chemical analysis of 18 pesticides, using gas chromatography with mass detector. ResultsWith passive air samplers we detected ten different pesticides: two insecticides, two nematicides, and six fungicides, of which nine reported to be used on banana plantations. More than half of the samples contained at least five pesticides. Chlorpyrifos was detected most-frequently, in 98% of samples, followed by the nematicides etoprophos and the fungicide pyrimethanil that were both detected in 81% of samples. Chlorpyrifos concentrations were five times higher in proximal as compared to non-proximal schools: mean = 18.2 ng/m3 (range = 6.1–36.1) and mean = 3.5 ng/m3 (range= <0.5–11.4) and varied more between schools than in time (intra-class correlation coefficient = 0.80). In general, results from passive and active samplers showed similar exposure patterns; yet median concentrations tended to be higher in passive samplers. In dust samples, mostly fungicides were detected; chlorothalonil was detected most frequently, in 50% of samples. DiscussionPassive air sampling is a promising technique to characterize environmental exposure to current-use pesticides; more studies are needed to characterize the sampling rates, reproducibility and optimum sampling times for passive samplers. School environments near banana plantations are contaminated with multiple pesticides that include insecticides, nematicides, and fungicides, which is of concern.

Atypical Repetition in Daily Conversation on Different Days for Detecting Alzheimer Disease: Evaluation of Phone-Call Data From Regular Monitoring Service.

BackgroundIdentifying signs of Alzheimer disease (AD) through longitudinal and passive monitoring techniques has become increasingly important. Previous studies have succeeded in quantifying language dysfunctions and identifying AD from speech data collected during neuropsychological tests. However, whether and how we can quantify language dysfunction in daily conversation remains unexplored.ObjectiveThe objective of this study was to explore the linguistic features that can be used for differentiating AD patients from daily conversations.MethodsWe analyzed daily conversational data of seniors with and without AD obtained from longitudinal follow-up in a regular monitoring service (from n=15 individuals including 2 AD patients at an average follow-up period of 16.1 months; 1032 conversational data items obtained during phone calls and approximately 221 person-hours). In addition to the standard linguistic features used in previous studies on connected speech data during neuropsychological tests, we extracted novel features related to atypical repetition of words and topics reported by previous observational and descriptive studies as one of the prominent characteristics in everyday conversations of AD patients.ResultsWhen we compared the discriminative power of AD, we found that atypical repetition in two conversations on different days outperformed other linguistic features used in previous studies on speech data during neuropsychological tests. It was also a better indicator than atypical repetition in single conversations as well as that in two conversations separated by a specific number of conversations.ConclusionsOur results show how linguistic features related to atypical repetition across days could be used for detecting AD from daily conversations in a passive manner by taking advantage of longitudinal data.