Close approximations to the sound of a cochlear implant.

Cochlear implant (CI) systems differ in terms of electrode design and signal processing. It is likely that patients fit with different implant systems will experience different percepts when presented speech via their implant. The sound quality of speech can be evaluated by asking single-sided-deaf (SSD) listeners fit with a cochlear implant (CI) to modify clean signals presented to their typically hearing ear to match the sound quality of signals presented to their CI ear. In this paper, we describe very close matches to CI sound quality, i.e., similarity ratings of 9.5 to 10 on a 10-point scale, by ten patients fit with a 28 mm electrode array and MED EL signal processing. The modifications required to make close approximations to CI sound quality fell into two groups: One consisted of a restricted frequency bandwidth and spectral smearing while a second was characterized by a wide bandwidth and no spectral smearing. Both sets of modifications were different from those found for patients with shorter electrode arrays who chose upshifts in voice pitch and formant frequencies to match CI sound quality. The data from matching-based metrics of CI sound quality document that speech sound-quality differs for patients fit with different CIs and among patients fit with the same CI.

A multi-resolution ensemble model of three decision-tree-based algorithms to predict daily NO2 concentration in France 2005-2022.

Understanding and managing the health effects of Nitrogen Dioxide (NO2) requires high resolution spatiotemporal exposure maps. Here, we developed a multi-stage multi-resolution ensemble model that predicts daily NO2 concentration across continental France from 2005 to 2022. Innovations of this work include the computation of daily predictions at a 200 m resolution in large urban areas and the use of a spatio-temporal blocking procedure to avoid data leakage and ensure fair performance estimation. Predictions were obtained after three cascading stages of modeling: (1) predicting NO2 total column density from Ozone Monitoring Instrument satellite; (2) predicting daily NO2 concentrations at a 1 km spatial resolution using a large set of potential predictors such as predictions obtained from stage 1, land-cover and road traffic data; and (3) predicting residuals from stage 2 models at a 200 m resolution in large urban areas. The latter two stages used a generalized additive model to ensemble predictions of three decision-tree algorithms (random forest, extreme gradient boosting and categorical boosting). Cross-validated performances of our ensemble models were overall very good, with a ten-fold cross-validated R2 for the 1 km model of 0.83, and of 0.69 for the 200 m model. All three basis learners participated in the ensemble predictions to various degrees depending on time and space. In sum, our multi-stage approach was able to predict daily NO2 concentrations with a relatively low error. Ensembling the predictions maximizes the chance of obtaining accurate values if one basis learner fails in a specific area or at a particular time, by relying on the other learners. To the best of our knowledge, this is the first study aiming to predict NO2 concentrations in France with such a high spatiotemporal resolution, large spatial extent, and long temporal coverage. Exposure estimates are available to investigate NO2 health effects in epidemiological studies.

Synchronizing Automatic Gain Control in Bilateral Cochlear Implants Mitigates Dynamic Localization Deficits Introduced by Independent Bilateral Compression.

OBJECTIVES: The independence of left and right automatic gain controls (AGCs) used in cochlear implants can distort interaural level differences and thereby compromise dynamic sound source localization. We assessed the degree to which synchronizing left and right AGCs mitigates those difficulties as indicated by listeners' ability to use the changes in interaural level differences that come with head movements to avoid front-back reversals (FBRs). DESIGN: Broadband noise stimuli were presented from one of six equally spaced loudspeakers surrounding the listener. Sound source identification was tested for stimuli presented at 70 dBA (above AGC threshold) for 10 bilateral cochlear implant patients, under conditions where (1) patients remained stationary and (2) free head movements within ±30° were encouraged. These conditions were repeated for both synchronized and independent AGCs. The same conditions were run at 50 dBA, below the AGC threshold, to assess listeners' baseline performance when AGCs were not engaged. In this way, the expected high variability in listener performance could be separated from effects of independent AGCs to reveal the degree to which synchronizing AGCs could restore localization performance to what it was without AGC compression. RESULTS: The mean rate of FBRs was higher for sound stimuli presented at 70 dBA with independent AGCs, both with and without head movements, than at 50 dBA, suggesting that when AGCs were independently engaged they contributed to poorer front-back localization. When listeners remained stationary, synchronizing AGCs did not significantly reduce the rate of FBRs. When AGCs were independent at 70 dBA, head movements did not have a significant effect on the rate of FBRs. Head movements did have a significant group effect on the rate of FBRs at 50 dBA when AGCs were not engaged and at 70 dBA when AGCs were synchronized. Synchronization of AGCs, together with head movements, reduced the rate of FBRs to approximately what it was in the 50-dBA baseline condition. Synchronizing AGCs also had a significant group effect on listeners' overall percent correct localization. CONCLUSIONS: Synchronizing AGCs allowed for listeners to mitigate front-back confusions introduced by unsynchronized AGCs when head motion was permitted, returning individual listener performance to roughly what it was in the 50-dBA baseline condition when AGCs were not engaged. Synchronization of AGCs did not overcome localization deficiencies which were observed when AGCs were not engaged, and which are therefore unrelated to AGC compression.

Spatial Heterogeneity Effects on Meta-Community Stability of Annual Plants from a Coastal Dune Ecosystem.

Spatial heterogeneity affects plant community composition and diversity. It is particularly noticeable in annual plant communities, which vary in space and time over short distances and periods, forming meta-communities at the regional scale. This study was conducted at the coastal dune ecosystem in Nizzanim nature reserve, Israel. This study aimed to analyze the effect of the spatial heterogeneity, which is expressed in differences in the fixation levels of the dunes and patches outside and beneath the dominant Artemisia monosperma shrubs, on the characteristics of the annual plant meta-community and its temporal stability, considering the mechanisms that may affect it. Thirteen dunes were studied: three mobile, seven semi-fixed, and three fixed dunes. Data on the annual plants were collected during the spring seasons of 2006, 2007, 2009, 2014, 2015, and 2016. For each dune, 72 quadrats of 40×40 cm were sampled yearly, with 24 quadrats per slope aspect (windward, leeward, and crest), 12 under the shrub, and 12 in the open. The results indicate that the transition from mobile dunes through semi-fixed to fixed dunes is characterized by an increase in annual plant cover, species richness, species diversity, changes in plant communities, and stability driven by the asynchrony of species population fluctuations. Asynchrony affected the stability of the meta-community of this ecosystem in patches beneath the shrubs but not in the open patches.

Predicting fine-scale daily NO2 over Mexico City using an ensemble modeling approach.

In recent years, there has been growing interest in developing air pollution prediction models to reduce exposure measurement error in epidemiologic studies. However, efforts for localized, fine-scale prediction models have been predominantly focused in the United States and Europe. Furthermore, the availability of new satellite instruments such as the TROPOsopheric Monitoring Instrument (TROPOMI) provides novel opportunities for modeling efforts. We estimated daily ground-level nitrogen dioxide (NO2) concentrations in the Mexico City Metropolitan Area at 1-km2 grids from 2005 to 2019 using a four-stage approach. In stage 1 (imputation stage), we imputed missing satellite NO2 column measurements from the Ozone Monitoring Instrument (OMI) and TROPOMI using the random forest (RF) approach. In stage 2 (calibration stage), we calibrated the association of column NO2 to ground-level NO2 using ground monitors and meteorological features using RF and extreme gradient boosting (XGBoost) models. In stage 3 (prediction stage), we predicted the stage 2 model over each 1-km2 grid in our study area, then ensembled the results using a generalized additive model (GAM). In stage 4 (residual stage), we used XGBoost to model the local component at the 200-m2 scale. The cross-validated R2 of the RF and XGBoost models in stage 2 were 0.75 and 0.86 respectively, and 0.87 for the ensembled GAM. Cross-validated rootmean-squared error (RMSE) of the GAM was 3.95 µg/m3. Using novel approaches and newly available remote sensing data, our multi-stage model presented high cross-validated fits and reconstructs fine-scale NO2 estimates for further epidemiologic studies in Mexico City.

A sensitive soil biological indicator to changes in land-use in regions with Mediterranean climate.

The demand for reliable indicators to quantify soil health has increased recently. We propose and test the use of soil microbial functional diversity as an indicator of multifunctional performance in agriculturally important areas. Agricultural fields in the Mediterranean and semiarid regions of Israel were selected as test sites and measured in Spring and Autumn seasons. Measurements included microbial parameters, basic soil abiotic properties and biological responses to agricultural management relative to measures of a natural ecosystem. Using a canonical correlation analysis we found that soil moisture was the most important basic soil property with different responses in Spring and Autumn. In Spring, it had a strongly negative relation with microbial biomass (MB), community level physiological profiling (CLPP) and the Shannon-Weaver index H', while in Autumn it had a strong relation with CLPP. We further show a significant interaction between CLPP and climate for land-use type "orchards". CLPP measured in the autumn season was thus identified as a useful and rapid biological soil health indicator, recommended for application in semiarid and Mediterranean agricultural regions. Apart from obtaining a better understanding of CLPP as the soil indicator, the study concludes that CLPP is well suited to differentiate between soils in different climates, seasons and land use types. The study shows a promising direction for further research on characterizing soil health under a larger variety of conditions.

AzBio Sentence test in Hebrew (HeBio): development, preliminary validation, and the effect of noise.

The AzBio sentence test is widely used to assess speech perception pre- and post-cochlear implantation. This study created and validated a Hebrew version of AzBio (HeBio) and tested its intelligibility amidst background noise.In Experiment 1, 1,000 recorded Hebrew sentences were presented via five-channel vocoder to 10 normal hearing (NH) listeners for intelligibility testing. In Experiment 2, HeBio lists were presented to 25 post-lingual cochlear implant (CI) users amidst four-talker babble noise (4TBN) or in quiet, along with one-syllable word test. In Experiment 3, 20 NH listeners were presented with eight HeBio lists in two noise conditions [4TBN, speech shaped noise (SSN)] and four SNRs (+3, 0 dB, -3 dB, -6 dB).HeBio lists (33) produced 82% average understanding, no inter-list intelligibility differences among NH, and equal intelligibility for CI users. One-syllable words predicted 67% of the variance in HeBio among CI users. Higher intelligibility was found for SSN than for 4TBN, and the mean speech receptive threshold (SRT) was more negative for SSN than for 4TBN.HeBio results were similar to AzBio. Results obtained with two noise types were as expected. HeBio is recommended for evaluation of different populations in quiet and noise.

Nature and happiness in an individualist and a collectivist culture.

According to the attention restoration theory, exposure to nature (ETN) renews one's capacity to focus attention, which decreases cognitive fatigue and therefore may increase positive emotions. Indeed, natural settings have been associated with high prevalence of happy facial expressions (HFE). However, how universal the association is, remains unclear. We explored the ETN-HFE association in Boston, US, representing a less collectivistic culture, and Yokohama, Japan, representing a more collectivistic one. Evidence from satellite images and social network data, using geoinformatics and statistical tools, revealed that individuals from both societies exhibited more happiness when they were photographed in more natural settings. These associations varied with temporal variations expressed through weekly and annual effects. In addition, we found that the presence of others was also associated with prevalence of HFE in natural settings at Yokohama and Boston but the relation was significantly stronger in Boston. Despite some relatively minor differences between the countries, these results support the universality of the association between ETN and HFE.

Upward Shifts in the Internal Representation of Frequency Can Persist Over a 3-Year Period for Cochlear Implant Patients Fit With a Relatively Short Electrode Array.

Patients fit with cochlear implants (CIs) commonly indicate at the time of device fitting and for some time after, that the speech signal sounds abnormal. A high pitch or timbre is one component of the abnormal percept. In this project, our aim was to determine whether a number of years of CI use reduced perceived upshifts in frequency spectrum and/or voice fundamental frequency. The participants were five individuals who were deaf in one ear and who had normal hearing in the other ear. The deafened ears had been implanted with a 18.5 mm electrode array which resulted in signal input frequencies being directed to locations in the spiral ganglion (SG) that were between one and two octaves higher than the input frequencies. The patients judged the similarity of a clean signal (a male-voice sentence) presented to their implanted ear and candidate, implant-like, signals presented to their normal-hearing (NH) ear. Matches to implant sound quality were obtained, on average, at 8 months after device activation (see section "Time 1") and at 35 months after activation (see section "Time 2"). At Time 1, the matches to CI sound quality were characterized, most generally, by upshifts in the frequency spectrum and in voice pitch. At Time 2, for four of the five patients, frequency spectrum values remained elevated. For all five patients F0 values remained elevated. Overall, the data offer little support for the proposition that, for patients fit with shorter electrode arrays, cortical plasticity nudges the cortical representation of the CI voice toward more normal, or less upshifted, frequency values between 8 and 35 months after device activation. Cortical plasticity may be limited when there are large differences between frequencies in the input signal and the locations in the SG stimulated by those frequencies.

Effects of Bilateral Automatic Gain Control Synchronization in Cochlear Implants With and Without Head Movements: Sound Source Localization in the Frontal Hemifield.

Purpose For bilaterally implanted patients, the automatic gain control (AGC) in both left and right cochlear implant (CI) processors is usually neither linked nor synchronized. At high AGC compression ratios, this lack of coordination between the two processors can distort interaural level differences, the only useful interaural difference cue available to CI patients. This study assessed the improvement, if any, in the utility of interaural level differences for sound source localization in the frontal hemifield when AGCs were synchronized versus independent and when listeners were stationary versus allowed to move their heads. Method Sound source identification of broadband noise stimuli was tested for seven bilateral CI patients using 13 loudspeakers in the frontal hemifield, under conditions where AGCs were linked and unlinked. For half the conditions, patients remained stationary; in the other half, they were encouraged to rotate or reorient their heads within a range of approximately ± 30° during sound presentation. Results In general, those listeners who already localized reasonably well with independent AGCs gained the least from AGC synchronization, perhaps because there was less room for improvement. Those listeners who performed worst with independent AGCs gained the most from synchronization. All listeners performed as well or better with synchronization than without; however, intersubject variability was high. Head movements had little impact on the effectiveness of synchronization of AGCs. Conclusion Synchronization of AGCs offers one promising strategy for improving localization performance in the frontal hemifield for bilaterally implanted CI patients. Supplemental Material https://doi.org/10.23641/asha.14681412.

A 1-km hourly air-temperature model for 13 northeastern U.S. states using remotely sensed and ground-based measurements.

BACKGROUND: Accurate and precise estimates of ambient air temperatures that can capture fine-scale within-day variability are necessary for studies of air temperature and health. METHOD: We developed statistical models to predict temperature at each hour in each cell of a 927-m square grid across the Northeast and Mid-Atlantic United States from 2003 to 2019, across ~4000 meteorological stations from the Integrated Mesonet, using inputs such as elevation, an inverse-distance-weighted interpolation of temperature, and satellite-based vegetation and land surface temperature. We used a rigorous spatial cross-validation scheme and spatially weighted the errors to estimate how well model predictions would generalize to new cell-days. We assess the within-county association of temperature and social vulnerability in a heat wave as an example application. RESULTS: We found that a model based on the XGBoost machine-learning algorithm was fast and accurate, obtaining weighted root mean square errors (RMSEs) around 1.6 K, compared to standard deviations around 11.0 K. We found similar accuracy when validating our model on an external dataset from Weather Underground. Assessing predictions from the North American Land Data Assimilation System-2 (NLDAS-2), another hourly model, in the same way, we found it was much less accurate, with RMSEs around 2.5 K. This is likely due to the NLDAS-2 model's coarser spatial resolution, and the dynamic variability of temperature within its grid cells. Finally, we demonstrated the health relevance of our model by showing that our temperature estimates were associated with social vulnerability across the region during a heat wave, whereas the NLDAS-2 showed a much weaker association. CONCLUSION: Our high spatiotemporal resolution air temperature model provides a strong contribution for future health studies in this region.

Development and Validation of the Spanish AzBio Sentence Corpus.

OBJECTIVE: To create and validate a Spanish sentence test for evaluation of speech understanding of Spanish-speaking listeners with hearing loss or cochlear implants (CI). STUDY DESIGN: Two thousand sentences were recorded from two male and two female speakers. The average intelligibility of each sentence was estimated as the mean score achieved by five listeners presented with a five-channel cochlear implant simulation. The mean scores of each sentence were used to construct 42 lists of 20 sentences with similar mean scores. List equivalency was then validated by presenting all lists to 10 CI users and in a 2-list comparison in a clinical setting to 38 CI patients. SETTING: Tertiary referral center. PATIENTS: Normal-hearing listeners (n = 5), CI users in a research study (n = 10), and CI patients (n = 38) in routine clinical follow-up. INTERVENTION: Multiple sentence lists from a newly minted speech perception test. MAIN OUTCOME MEASURES: List intelligibility and equivalence across sentence lists. RESULTS: Forty-two lists of sentences were equivalent when all lists were presented in random order to 10 adult CI recipients. The variability of scores observed on lists presented to the same listener in the same condition was captured using a binomial distribution model based on a 40-item list for 38 adult implant recipients. CONCLUSION: The Spanish AzBio Sentence Test includes 42 lists of 20 sentences. These sentences are roughly equivalent in terms of overall difficulty and confidence limits have been provided to assess the significance of variability in list scores observed within or across conditions. These materials will be of benefit when assessing native Spanish speakers in both research and clinical settings.

Impact of Botulinum Toxin on Hirschsprung-Associated Enterocolitis After Primary Pull-Through.

BACKGROUND: Hirschsprung-associated enterocolitis (HAEC) is a serious potential complication after primary pull-through surgery for Hirschsprung's disease (HSCR). Administration of anal botulinum toxin (BT) injection may improve obstructive symptoms at the internal anal sphincter, leading to improved fecal passage. The timing of administration and effects on delay or prevention of HAEC are unknown. We hypothesized that BT administration increased the postoperative time to HAEC and aimed to investigate whether anal BT administration after primary pull-through surgery for HSCR is associated with increased time to inpatient HAEC admission development. METHODS: We performed a retrospective cohort study examining children with HSCR at US children's hospitals from 2008 to 2018 using the Pediatric Health Information System database with an associated primary pull-through operation performed before 60 d of age. The intervention assessed was the administration of BT concerning the timing of primary pull-through, and two groups were identified: PRO (received BT at or after primary pull-through, before HAEC) and NOT (never received BT, or received BT after HAEC). The primary outcome was time from pull-through to the first HAEC admission. The Cox proportional hazards model was developed to examine the BT administration effect on the primary outcome after controlling for patient-level covariates. RESULTS: We examined a total of 1439 children (67 in the PRO and 1372 in the NOT groups). A total of 308 (21.4%) developed at least one episode of HAEC, including 76 (5.3%) who had two or more episodes. Between 2008 and 2018, the frequency of BT administration has increased from three to 20 hospitals with a frequency of administration between 2.2% and 16.2%. Prophylactic BT (PRO) was not associated with increased time to HAEC event on adjusted analysis. CONCLUSIONS: Among children with HSCR undergoing primary pull-through surgery, prophylactic BT administration did not demonstrate increased time to first HAEC event. A better-powered study with prophylactic BT is required to determine the effect on HAEC occurrence and timing. LEVEL OF EVIDENCE: Level II (retrospective cohort study).

The Benefit of Remote and On-Ear Directional Microphone Technology Persists in the Presence of Visual Information.

BACKGROUND: Both the Roger remote microphone and on-ear, adaptive beamforming technologies (e.g., Phonak UltraZoom) have been shown to improve speech understanding in noise for cochlear implant (CI) listeners when tested in audio-only (A-only) test environments. PURPOSE: Our aim was to determine if adult and pediatric CI recipients benefited from these technologies in a more common environment-one in which both audio and visual cues were available and when overall performance was high. STUDY SAMPLE: Ten adult CI listeners (Experiment 1) and seven pediatric CI listeners (Experiment 2) were tested. DESIGN: Adults were tested in quiet and in two levels of noise (level 1 and level 2) in A-only and audio-visual (AV) environments. There were four device conditions: (1) an ear canal-level, omnidirectional microphone (T-mic) in quiet, (2) the T-mic in noise, (3) an adaptive directional mic (UltraZoom) in noise, and (4) a wireless, remote mic (Roger Pen) in noise. Pediatric listeners were tested in quiet and in level 1 noise in A-only and AV environments. The test conditions were: (1) a behind-the-ear level omnidirectional mic (processor mic) in quiet, (2) the processor mic in noise, (3) the T-mic in noise, and (4) the Roger Pen in noise. DATA COLLECTION AND ANALYSES: In each test condition, sentence understanding was assessed (percent correct) and ease of listening ratings were obtained. The sentence understanding data were entered into repeated-measures analyses of variance. RESULTS: For both adult and pediatric listeners in the AV test conditions in level 1 noise, performance with the Roger Pen was significantly higher than with the T-mic. For both populations, performance in level 1 noise with the Roger Pen approached the level of baseline performance in quiet. Ease of listening in noise was rated higher in the Roger Pen conditions than in the T-mic or processor mic conditions in both A-only and AV test conditions. CONCLUSION: The Roger remote mic and on-ear directional mic technologies benefit both speech understanding and ease of listening in a realistic laboratory test environment and are likely do the same in real-world listening environments.

Validating and Comparing Highly Resolved Commercial "Off the Shelf" PM Monitoring Sensors with Satellite Based Hybrid Models, for Improved Environmental Exposure Assessment.

Particulate matter is a common health hazard, and under certain conditions, an ecological threat. While many studies were conducted in regard to air pollution and potential effects, this paper serves as a pilot scale investigation into the spatial and temporal variability of particulate matter (PM) pollution in arid urban environments in general, and Beer-Sheva, Israel as a case study. We explore the use of commercially off the shelf (COTS) sensors, which provide an economical solution for spatio-temporal measurements. We started with a comparison process against an A-grade meteorological station, where it was shown that under specific climatic conditions, a number of COTS sensors were able to produce robust agreement (mean R2=0.93, average SD=17.5). The second stage examined the COTS sensors that were proven accurate in a mobile measurement campaign. Finally, data collected was compared to a validated satellite prediction model. We present how these tests and COTS sensor-kits could then be used to further explain the continuity and dispersion of particulate matter in similar areas.

Effects of Head Movements on Sound-Source Localization in Single-Sided Deaf Patients With Their Cochlear Implant On Versus Off.

OBJECTIVES: We investigated the ability of single-sided deaf listeners implanted with a cochlear implant (SSD-CI) to (1) determine the front-back and left-right location of sound sources presented from loudspeakers surrounding the listener and (2) use small head rotations to further improve their localization performance. The resulting behavioral data were used for further analyses investigating the value of so-called "monaural" spectral shape cues for front-back sound source localization. DESIGN: Eight SSD-CI patients were tested with their cochlear implant (CI) on and off. Eight normal-hearing (NH) listeners, with one ear plugged during the experiment, and another group of eight NH listeners, with neither ear plugged, were also tested. Gaussian noises of 3-sec duration were band-pass filtered to 2-8 kHz and presented from 1 of 6 loudspeakers surrounding the listener, spaced 60° apart. Perceived sound source localization was tested under conditions where the patients faced forward with the head stationary, and under conditions where they rotated their heads between (Equation is included in full-text article.). RESULTS: (1) Under stationary listener conditions, unilaterally-plugged NH listeners and SSD-CI listeners (with their CIs both on and off) were nearly at chance in determining the front-back location of high-frequency sound sources. (2) Allowing rotational head movements improved performance in both the front-back and left-right dimensions for all listeners. (3) For SSD-CI patients with their CI turned off, head rotations substantially reduced front-back reversals, and the combination of turning on the CI with head rotations led to near-perfect resolution of front-back sound source location. (4) Turning on the CI also improved left-right localization performance. (5) As expected, NH listeners with both ears unplugged localized to the correct front-back and left-right hemifields both with and without head movements. CONCLUSIONS: Although SSD-CI listeners demonstrate a relatively poor ability to distinguish the front-back location of sound sources when their head is stationary, their performance is substantially improved with head movements. Most of this improvement occurs when the CI is off, suggesting that the NH ear does most of the "work" in this regard, though some additional gain is introduced with turning the CI on. During head turns, these listeners appear to primarily rely on comparing changes in head position to changes in monaural level cues produced by the direction-dependent attenuation of high-frequency sounds that result from acoustic head shadowing. In this way, SSD-CI listeners overcome limitations to the reliability of monaural spectral and level cues under stationary conditions. SSD-CI listeners may have learned, through chronic monaural experience before CI implantation, or with the relatively impoverished spatial cues provided by their CI-implanted ear, to exploit the monaural level cue. Unilaterally-plugged NH listeners were also able to use this cue during the experiment to realize approximately the same magnitude of benefit from head turns just minutes after plugging, though their performance was less accurate than that of the SSD-CI listeners, both with and without their CI turned on.

Gradient boosting machine learning to improve satellite-derived column water vapor measurement error.

The atmospheric products of the Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm include column water vapor (CWV) at a 1 km resolution, derived from daily overpasses of NASA's Moderate Resolution Imaging Spectroradiometer (MODIS) instruments aboard the Aqua and Terra satellites. We have recently shown that machine learning using extreme gradient boosting (XGBoost) can improve the estimation of MAIAC aerosol optical depth (AOD). Although MAIAC CWV is generally well validated (Pearson's R >0.97 versus CWV from AERONET sun photometers), it has not yet been assessed whether machine-learning approaches can further improve CWV. Using a novel spatiotemporal cross-validation approach to avoid overfitting, our XGBoost model, with nine features derived from land use terms, date, and ancillary variables from the MAIAC retrieval, quantifies and can correct a substantial portion of measurement error relative to collocated measurements at AERONET sites (26.9% and 16.5% decrease in root mean square error (RMSE) for Terra and Aqua datasets, respectively) in the Northeastern USA, 2000-2015. We use machine-learning interpretation tools to illustrate complex patterns of measurement error and describe a positive bias in MAIAC Terra CWV worsening in recent summertime conditions. We validate our predictive model on MAIAC CWV estimates at independent stations from the SuomiNet GPS network where our corrections decrease the RMSE by 19.7% and 9.5% for Terra and Aqua MAIAC CWV. Empirically correcting for measurement error with machine-learning algorithms is a postprocessing opportunity to improve satellite-derived CWV data for Earth science and remote sensing applications.

Advancing methodologies for applying machine learning and evaluating spatiotemporal models of fine particulate matter (PM2.5) using satellite data over large regions.

Reconstructing the distribution of fine particulate matter (PM2.5) in space and time, even far from ground monitoring sites, is an important exposure science contribution to epidemiologic analyses of PM2.5 health impacts. Flexible statistical methods for prediction have demonstrated the integration of satellite observations with other predictors, yet these algorithms are susceptible to overfitting the spatiotemporal structure of the training datasets. We present a new approach for predicting PM2.5 using machine-learning methods and evaluating prediction models for the goal of making predictions where they were not previously available. We apply extreme gradient boosting (XGBoost) modeling to predict daily PM2.5 on a 1×1 km2 resolution for a 13 state region in the Northeastern USA for the years 2000-2015 using satellite-derived aerosol optical depth and implement a recursive feature selection to develop a parsimonious model. We demonstrate excellent predictions of withheld observations but also contrast an RMSE of 3.11 µg/m3 in our spatial cross-validation withholding nearby sites versus an overfit RMSE of 2.10 µg/m3 using a more conventional random ten-fold splitting of the dataset. As the field of exposure science moves forward with the use of advanced machine-learning approaches for spatiotemporal modeling of air pollutants, our results show the importance of addressing data leakage in training, overfitting to spatiotemporal structure, and the impact of the predominance of ground monitoring sites in dense urban sub-networks on model evaluation. The strengths of our resultant modeling approach for exposure in epidemiologic studies of PM2.5 include improved efficiency, parsimony, and interpretability with robust validation while still accommodating complex spatiotemporal relationships.

Associations between daily ambient temperature and sedentary time among children 4-6 years old in Mexico City.

BACKGROUND: Sedentary behavior is a worldwide public health concern. There is consistent and growing evidence linking sedentary behavior to mortality and morbidity. Early monitoring and assessment of environmental factors associated with sedentary behaviors at a young age are important initial steps for understanding children's sedentary time and identifying pertinent interventions. OBJECTIVE: This study examines the association between daily temperature (maximum, mean, minimum, and diurnal variation) and all-day sedentary time among 4-6 year old children in Mexico City (n = 559) from the year 2013 to 2015. METHODS: We developed a spatiotemporally resolved hybrid satellite-based land use regression temperature model and calculated percent daily sedentary time from aggregating 10-second epoch vertical counts captured by accelerometers that participants wore for one week. We modeled generalized additive models (GAMs), one for each temperature type as a covariate (maximum, mean, minimum, and diurnal variation). All GAMs included percent all-day sedentary time as the outcome and participant-level random intercepts to account for repeated measures of sedentary time. Our models were adjusted for demographic factors and environmental exposures. RESULTS: Daily maximum temperature, mean temperature, and diurnal variation have significant negative linear relationships with all-day sedentary time (p<0.01). There is no significant association between daily minimum temperature and all-day sedentary time. Children have on average 0.26% less daily sedentary time (approximately 2.2 minutes) for each 1°C increase in ambient maximum temperature (range 7.1-30.2°C), 0.27% less daily sedentary time (approximately 2.3 minutes) for each 1°C increase in ambient mean temperature (range 4.3-22.2°C), and 0.23% less daily sedentary time (approximately 2.0 minutes) for each 1°C increase in diurnal variation (range 3.0-21.6°C). CONCLUSIONS: These results are contrary to our hypothesis in which we expected a curvilinear relationship between temperature (maximum, mean, minimum, and diurnal variation) and sedentary time. Our findings suggest that temperature is an important environmental factor that influences children's sedentary behavior.

Spatiotemporal distribution of autism spectrum disorder prevalence among birth cohorts during 2000-2011 in Israel.

PURPOSE: Studies indicate an apparent sharp increasing trend in autism spectrum disorder (ASD) incidence and prevalence worldwide. This nationwide study aims at depicting ASD prevalence distribution in Israel in both space and time. METHODS: Based on data from Israel National Insurance Institute, the study population included all children born in Israel 2000-2011 (n = 1,786,194), of whom 11,699 (0.655%) were subsequently diagnosed with ASD (until December 31, 2016). Prevalence was calculated and mapped by dividing the number of ASD cases within each year of birth by the number of births during that year, for each spatial unit, and similarly for several spatiotemporal levels of aggregation. RESULTS: ASD prevalence varies substantially across different geographic areas in Israel, with considerably higher prevalence concentrated in central Israel. Strong associations were found between locality-level socioeconomic index, ethnicity, and peripherality and ASD prevalence, and even after adjustment for them, excess prevalence for ASD still persisted in certain localities. No spatial dependence of prevalence, with and without adjustment for the locality-level variables, was found (Moran's I = -0.000546, -0.00335, respectively). CONCLUSIONS: Our findings provide important insights regarding health disparities affecting ASD diagnosis, directing further health policy intervention and further research.