The value of deep learning-based computer aided diagnostic system in improving diagnostic performance of rib fractures in acute blunt trauma.

BACKGROUND: To evaluate the value of a deep learning-based computer-aided diagnostic system (DL-CAD) in improving the diagnostic performance of acute rib fractures in patients with chest trauma. MATERIALS AND METHODS: CT images of 214 patients with acute blunt chest trauma were retrospectively analyzed by two interns and two attending radiologists independently firstly and then with the assistance of a DL-CAD one month later, in a blinded and randomized manner. The consensusdiagnosis of fib fracture by another two senior thoracic radiologists was regarded as reference standard. The rib fracture diagnostic sensitivity, specificity, positive predictive value, diagnostic confidence and mean reading time with and without DL-CAD were calculated and compared. RESULTS: There were 680 rib fracture lesions confirmed as reference standard among all patients. The diagnostic sensitivity and positive predictive value of interns weresignificantly improved from (68.82%, 84.50%) to (91.76%, 93.17%) with the assistance of DL-CAD, respectively. Diagnostic sensitivity and positive predictive value of attendings aided by DL-CAD (94.56%, 95.67%) or not aided (86.47%, 93.83%), respectively. In addition, when radiologists were assisted by DL-CAD, the mean reading time was significantly reduced, and diagnostic confidence was significantly enhanced. CONCLUSIONS: DL-CAD improves the diagnostic performance of acute rib fracture in chest trauma patients, which increases the diagnostic confidence, sensitivity, and positive predictive value for radiologists. DL-CAD can advance the diagnostic consistency of radiologists with different experiences.

The Natural Stories corpus: a reading-time corpus of English texts containing rare syntactic constructions.

It is now a common practice to compare models of human language processing by comparing how well they predict behavioral and neural measures of processing difficulty, such as reading times, on corpora of rich naturalistic linguistic materials. However, many of these corpora, which are based on naturally-occurring text, do not contain many of the low-frequency syntactic constructions that are often required to distinguish between processing theories. Here we describe a new corpus consisting of English texts edited to contain many low-frequency syntactic constructions while still sounding fluent to native speakers. The corpus is annotated with hand-corrected Penn Treebank-style parse trees and includes self-paced reading time data and aligned audio recordings. We give an overview of the content of the corpus, review recent work using the corpus, and release the data.

Evaluation of the sensitivity of the Express View function in the Mirocam® capsule endoscopy software.

Background: A new computer algorithm called Express-View has recently been introduced by Mirocam, but data concerning its application and efficacy are scarce.Objective: To evaluate the lesion detection rate, per-patient sensitivity and the diagnostic accuracy using Express-View.Methods: All patients who performed CE between January 2018 and June 2019, whose indication was obscure gastrointestinal bleeding (OGIB) and with findings on CE, were included. Lesions identified in conventional reading were selected and considered as reference.Results: Eighty-nine patients were included, 50.6% male, with a mean age of 68.4 years-old (±12.3). The Express-View mode detected 85.5% of lesions previously detected by conventional reading (524 out of 613). There were 89 missed lesions, mainly erosions or ulcers (44.9%) and angioectasias (38.2%). The lesion detection rate was found to be lower in the jejunum and ileum compared to extra-small bowel locations and duodenum (p = .04). Although Express-View had a per-patient sensitivity for all lesions of 56.2% and a per-patient sensitivity for all clinically significant lesions of 83.1%, it achieved a diagnostic accuracy of 91%.Conclusions: The per-patient sensitivity for all lesions was shown to be below expectations, although the lesion detection rate, the per-patient sensitivity for all clinically significant lesions, and the diagnostic accuracy were shown to be higher.

Clinical usefulness of a deep learning-based system as the first screening on small-bowel capsule endoscopy reading.

BACKGROUND AND AIM: To examine whether our convolutional neural network (CNN) system based on deep learning can reduce the reading time of endoscopists without oversight of abnormalities in the capsule-endoscopy reading process. METHODS: Twenty videos of the entire small-bowel capsule endoscopy procedure were prepared, each of which included 0-5 lesions of small-bowel mucosal breaks (erosions or ulcerations). At another institute, two reading processes were compared: (A) endoscopist-alone readings and (B) endoscopist readings after the first screening by the proposed CNN. In process B, endoscopists read only images detected by CNN. Two experts and four trainees independently read 20 videos each (10 for process A and 10 for process B). Outcomes were reading time and detection rate of mucosal breaks by endoscopists. Gold standard was findings at the original institute by two experts. RESULTS: Mean reading time of small-bowel sections by endoscopists was significantly shorter during process B (expert, 3.1 min; trainee, 5.2 min) compared to process A (expert, 12.2 min; trainee, 20.7 min) (P < 0.001). For 37 mucosal breaks, detection rate by endoscopists did not significantly decrease in process B (expert, 87%; trainee, 55%) compared to process A (expert, 84%; trainee, 47%). Experts detected all eight large lesions (>5 mm), but trainees could not, even when supported by the CNN. CONCLUSIONS: Our CNN-based system for capsule endoscopy videos reduced the reading time of endoscopists without decreasing the detection rate of mucosal breaks. However, the reading level of endoscopists should be considered when using the system.

The impact of maternal perinatal depression on exposure to reading and screen time for their infants: pilot findings from the MPEWS Study.

OBJECTIVE: To explore the association between maternal depression and the screen and reading time experienced by their infants. METHODS: This study utilises data on 158 women and infants, collected within the Mercy Pregnancy and Emotional Wellbeing Study. Women less than 20 weeks gestation were diagnosed using the Structured Clinical Interview for DSM-IV Axis I Disorders. Six months postpartum they completed questionnaires about themselves, their infant and early parenting practices. RESULTS: Children of women with a past diagnosis of depression were exposed to fewer days of 15-minute reading time per week compared to the children of women with no diagnosis. While the current depression group showed a lower average reading time, this difference was not statistically significant. There were no significant differences in infant screen time between groups. CONCLUSIONS: A maternal past diagnosis of depression is correlated with decreased reading time in infants. This may present a practical point for screening and intervention or suggest a causal pathway for poorer outcomes in children of those with depression.

Accuracy and reading time for six strategies using digital breast tomosynthesis in women with mammographically negative dense breasts.

OBJECTIVE: To compare six strategies using digital breast tomosynthesis in women with mammographically negative dense breasts. MATERIALS AND METHODS: This is a substudy of the 'ASTOUND' trial. 163 women who underwent tomosynthesis with synthetically reconstructed projection images (S-2D) inclusive of 13 (7.9%) cases diagnosed with breast cancer at histopathology after surgery were evaluated. Accuracy measures and screen-reading time of six reading strategies were assessed: (A) Single reading of S-2D alone, (B) single reading of tomosynthesis alone, (C) single reading of joint interpretation of tomosynthesis + S-2D, (D) double-reading of S-2D alone, (E) double reading of tomosynthesis alone, (F) double reading of joint interpretation of tomosynthesis + S-2D. RESULTS: The median age of the patients was 53 years (range, 36-88 years). The highest global accuracy was obtained with double reading of tomosynthesis + S2D (F) with an AUC of 0.979 (p<0.001) and a mean reading time of 154 s versus 34 s for the fastest strategy (single reading of S-2D alone). The AUCs for the other five strategies did not differ from each other. CONCLUSION: Double reading of tomosynthesis+ S2D had the best accuracy of six screen-reading strategies although it had the longest reading time. KEY POINTS: • Tomosynthesis acquisitions are progressively implemented with reconstructed synthesized 2D images • Double reading using S-2D plus tomosynthesis had the highest global accuracy (p<0.001). • Double reading of S-2D plus tomosynthesis increased reading time.

Tumor Response Evaluation Using iRECIST: Feasibility and Reliability of Manual Versus Software-Assisted Assessments.

OBJECTIVES: To compare the feasibility and reliability of manual versus software-assisted assessments of computed tomography scans according to iRECIST in patients undergoing immune-based cancer treatment. METHODS: Computed tomography scans of 30 tumor patients undergoing cancer treatment were evaluated by four independent radiologists at baseline (BL) and two follow-ups (FU), resulting in a total of 360 tumor assessments (120 each at BL/FU1/FU2). After image interpretation, tumor burden and response status were either calculated manually or semi-automatically as defined by software, respectively. The reading time, calculated sum of longest diameter (SLD), and tumor response (e.g., "iStable Disease") were determined for each assessment. After complete data collection, a consensus reading among the four readers was performed to establish a reference standard for the correct response assignments. The reading times, error rates, and inter-reader agreement on SLDs were statistically compared between the manual versus software-assisted approaches. RESULTS: The reading time was significantly longer for the manual versus software-assisted assessments at both follow-ups (median [interquartile range] FU1: 4.00 min [2.17 min] vs. 2.50 min [1.00 min]; FU2: 3.75 min [1.88 min] vs. 2.00 min [1.50 min]; both p < 0.001). Regarding reliability, 2.5% of all the response assessments were incorrect at FU1 (3.3% manual; 0% software-assisted), which increased to 5.8% at FU2 (10% manual; 1.7% software-assisted), demonstrating higher error rates for manual readings. Quantitative SLD inter-reader agreement was inferior for the manual compared to the software-assisted assessments at both FUs (FU1: ICC = 0.91 vs. 0.93; FU2: ICC = 0.75 vs. 0.86). CONCLUSIONS: Software-assisted assessments may facilitate the iRECIST response evaluation of cancer patients in clinical routine by decreasing the reading time and reducing response misclassifications.

New Generation Express View: An Artificial Intelligence Software Effectively Reduces Capsule Endoscopy Reading Times.

BACKGROUND: Reading capsule endoscopy (CE) is time-consuming. The Express View (EV) (IntroMedic, Seoul, Korea) software was designed to shorten CE video reading. Our primary aim was to evaluate the diagnostic accuracy of EV in detecting significant small-bowel (SB) lesions. We also compared the reading times with EV mode and standard reading (SR). METHODS: 126 patients with suspected SB bleeding and/or suspected neoplasia were prospectively enrolled and underwent SB CE (MiroCam®1200, IntroMedic, Seoul, Korea). CE evaluation was performed in standard and EV mode. In case of discrepancies between SR and EV readings, a consensus was reached after reviewing the video segments and the findings were re-classified. RESULTS: The completion rate of SB CE in our cohort was 86.5% and no retention occurred. The per-patient analysis of sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy of EV compared to SR were 86%, 86%, 90%, 81%, and 86%, respectively, before consensus. After consensus, they increased to 97%, 100%, 100%, 96%, and 98%, respectively. The median reading time with SR and EV was 71 min (range 26−340) and 13 min (range 3−85), respectively (p < 0.001). CONCLUSIONS: The new-generation EV shows high diagnostic accuracy and significantly reduces CE reading times.

Improving the Performance of Radiologists Using Artificial Intelligence-Based Detection Support Software for Mammography: A Multi-Reader Study.

OBJECTIVE: To evaluate whether artificial intelligence (AI) for detecting breast cancer on mammography can improve the performance and time efficiency of radiologists reading mammograms. MATERIALS AND METHODS: A commercial deep learning-based software for mammography was validated using external data collected from 200 patients, 100 each with and without breast cancer (40 with benign lesions and 60 without lesions) from one hospital. Ten readers, including five breast specialist radiologists (BSRs) and five general radiologists (GRs), assessed all mammography images using a seven-point scale to rate the likelihood of malignancy in two sessions, with and without the aid of the AI-based software, and the reading time was automatically recorded using a web-based reporting system. Two reading sessions were conducted with a two-month washout period in between. Differences in the area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, and reading time between reading with and without AI were analyzed, accounting for data clustering by readers when indicated. RESULTS: The AUROC of the AI alone, BSR (average across five readers), and GR (average across five readers) groups was 0.915 (95% confidence interval, 0.876-0.954), 0.813 (0.756-0.870), and 0.684 (0.616-0.752), respectively. With AI assistance, the AUROC significantly increased to 0.884 (0.840-0.928) and 0.833 (0.779-0.887) in the BSR and GR groups, respectively (p = 0.007 and p < 0.001, respectively). Sensitivity was improved by AI assistance in both groups (74.6% vs. 88.6% in BSR, p < 0.001; 52.1% vs. 79.4% in GR, p < 0.001), but the specificity did not differ significantly (66.6% vs. 66.4% in BSR, p = 0.238; 70.8% vs. 70.0% in GR, p = 0.689). The average reading time pooled across readers was significantly decreased by AI assistance for BSRs (82.73 vs. 73.04 seconds, p < 0.001) but increased in GRs (35.44 vs. 42.52 seconds, p < 0.001). CONCLUSION: AI-based software improved the performance of radiologists regardless of their experience and affected the reading time.

What's new? A comprehension bias in favor of informativity.

Language is used as a channel by which speakers convey, among other things, newsworthy and informative messages, i.e., content that is otherwise unpredictable to the comprehender. We therefore might expect comprehenders to show a preference for such messages. However, comprehension studies tend to emphasize the opposite: i.e., processing ease for situation-predictable content (e.g., chopping carrots with a knife). Comprehenders are known to deploy knowledge about situation plausibility during processing in fine-grained context-sensitive ways. Using self-paced reading, we test whether comprehenders can also deploy this knowledge in favor of newsworthy content to yield informativity-driven effects alongside, or instead of, plausibility-driven effects. We manipulate semantic context (unusual protagonists), syntactic construction (wh- clefts), and the communicative environment (text messages). Reading times (primarily sentence-finally) show facilitation for sentences containing newsworthy content (e.g., chopping carrots with a shovel), where the content is both unpredictable at the situation level because of its atypicality and also unpredictable at the word level because of the large number of atypical elements a speaker could potentially mention. Our studies are the first to show that informativity-driven effects are observable at all, and the results highlight the need for models that distinguish between comprehenders' estimate of content plausibility and their estimate of a speaker's decision to talk about that content.

Optimal clinical time for reliable measurement of transcutaneous CO2 with ear probes: counterbalancing overshoot and the vasodilatation effect.

OBJECTIVES: To determine the optimal clinical reading time for the transcutaneous measurement of oxygen saturation (SpO(2)) and transcutaneous CO(2) (TcPCO(2)) in awake spontaneously breathing individuals, considering the overshoot phenomenon (transient overestimation of arterial PaCO(2)). EXPERIMENTAL SECTION: Observational study of 91 (75 men) individuals undergoing forced spirometry, measurement of SpO(2) and TcPCO(2) with the SenTec monitor every two minutes until minute 20 and arterial blood gas (ABG) analysis. Overshoot severity: (a) mild (0.1-1.9 mm Hg); (b) moderate (2-4.9 mm Hg); (c) severe: (>5 mm Hg). The mean difference was calculated for SpO(2) and TcPCO(2) and arterial values of PaCO(2) and SpO(2). The intraclass correlation coefficient (ICC) between monitor readings and blood values was calculated as a measure of agreement. RESULTS: The mean age was 63.1 ± 11.8 years. Spirometric values: FVC: 75.4 ± 6.2%; FEV(1): 72.9 ± 23.9%; FEV(1)/FVC: 70 ± 15.5%. ABG: PaO(2): 82.6 ± 13.2; PaCO(2): 39.9.1 ± 4.8 mmHg; SaO(2): 95.3 ± 4.4%. Overshoot analysis: overshoot was mild in 33 (36.3%) patients, moderate in 20 (22%) and severe in nine (10%); no overshoot was observed in 29 (31%) patients. The lowest mean differences between arterial blood gas and TcPCO(2) was -0.57 mmHg at minute 10, although the highest ICC was obtained at minutes 12 and 14 (>0.8). The overshoot lost its influence after minute 12. For SpO(2), measurements were reliable at minute 2. CONCLUSIONS: The optimal clinical reading measurement recommended for the ear lobe TcPCO(2) measurement ranges between minute 12 and 14. The SpO(2) measurement can be performed at minute 2.

Reading Fast, Reading Slow: The Effect of Interviewers' Speed in Reading Introductory Texts on Response Behavior.

Guidelines for interviewers frequently include instructions to read question texts exactly as they are worded. Deviations from these guidelines on standardized interviewing might affect the comparability of survey answers and impair the quality of data. This paper contributes to the literature on interviewer behavior by analyzing how interviewers change their reading behavior during fieldwork and whether this behavioral change influences the response behavior of survey respondents. We use item-level paradata from the Survey of Health, Ageing and Retirement in Europe (SHARE) to measure interviewers' reading times and focus our analyses on introductory questions that do not require an immediate response by the respondent. In contrast to prior research, this focus enables us to disentangle the reading times of interviewers from the response times of respondents. Based on fixed effects regression, our results show systematic changes in interviewers' reading times of introductory items: First, reading times significantly decrease over the survey's field period, even after controlling for relevant respondent characteristics and specific aspects of the interview situation. Second, a cross-national comparison, including fourteen European countries plus Israel, reveals that the decrease is uniform in almost all countries, suggesting its generalizability across different cultural contexts. Third, the decrease in reading times influences response behavior to varying degrees. Response behavior is affected if introductions contain relevant information for understanding or fulfilling the required task and especially if the response refers to within-survey requests. On the basis of these findings, we discuss the possible consequences for questionnaire design, interviewer training, and fieldwork monitoring.

Evaluating information-theoretic measures of word prediction in naturalistic sentence reading.

We review information-theoretic measures of cognitive load during sentence processing that have been used to quantify word prediction effort. Two such measures, surprisal and next-word entropy, suffer from shortcomings when employed for a predictive processing view. We propose a novel metric, lookahead information gain, that can overcome these short-comings. We estimate the different measures using probabilistic language models. Subsequently, we put them to the test by analysing how well the estimated measures predict human processing effort in three data sets of naturalistic sentence reading. Our results replicate the well known effect of surprisal on word reading effort, but do not indicate a role of next-word entropy or lookahead information gain. Our computational results suggest that, in a predictive processing system, the costs of predicting may outweigh the gains. This idea poses a potential limit to the value of a predictive mechanism for the processing of language. The result illustrates the unresolved problem of finding estimations of word-by-word prediction that, first, are truly independent of perceptual processing of the to-be-predicted words, second, are statistically reliable predictors of experimental data, and third, can be derived from more general assumptions about the cognitive processes involved.

When Words and Graphs Move the Eyes: The Processing of Multimodal Causal Relations.

Research on causal relations in multisemiotic texts constituted by words and graphs has been scarce with only a few exceptions. In the current study, eye movement behavior was studied in seventy-six Chilean high school students, who read a set of twelve causally-related economics texts in Spanish in four experimental conditions. The objective is twofold. We aimed, on the one hand, to observe the main effects of the causal discourse marker (DM) por tanto and the statistical causal graph (G), as well as the interaction effect of both variables on different eye tracking measures. On the other, we seek to observe the effects of the DM on the same eye tracking measures for the graph system (GS) area of interest (AOI). The findings showed that the conjoint presence of the DM and the G did not positively influence the processing of selected AOIs. Analyses also reveal no significant effects on the GS AOI. Thus, the results indicate that the DM tend to decrease processing times, while the G increases them. Additional analyses conducted on the integrative transitions between the verbal system and the graph system reveal that more transitions were identified between the consequence segment and the graph system, thereby confirming that the consequence segment is crucial for the integration of both semiotic systems.

Correlations between working memory capacity and EFL reading comprehension: an investigation with Brazilian university students / Correlações entre capacidade de memória de trabalho e compreensão leitora em Inglês como Língua Estrangeira: uma investigação com estudantes universitários brasileiros

Reading in a foreign language is considered a high cognitive function, believed, thus, to impose demands on working memory resources. Studies have shown significant correlations between working memory capacity and foreign language reading comprehension, but its underlying mechanisms are still open to discussion. This article reports on a study whose objective was to investigate possible correlations between working memory capacity and foreign language reading comprehension considering the knowledge of the first language as a possible source of interference. A recall task was conducted in two conditions, one in which readers performed the task in English and another, in Portuguese. Results showed a weak but statistically significant correlation between working memory capacity and recall scores in the second condition and no correlations between working memory capacity and reading time. Exploratory data analyses suggested that the correlation observed was influenced by foreign language reading proficiency and that the ratio of recall scores to reading time is a better measure of comprehension efficiency.

A leitura em língua estrangeira é considerada uma função cognitiva de alto nível que se acredita, portanto, impor demandas nos recursos da memória de trabalho. Estudos mostram correlações significativas entre capacidade de memória de trabalho e compreensão leitora em língua estrangeira, mas seus mecanismos subjacentes ainda estão abertos a discussões. Esse artigo reporta um estudo cujo objetivo foi investigar possíveis correlações entre capacidade de memória de trabalho e compreensão leitora em língua estrangeira considerando o conhecimento da primeira língua como uma possível fonte de interferência. Uma tarefa de recordação foi conduzida em duas condições, uma em que os leitores realizaram a tarefa em inglês e outra, em português. Os resultados mostraram uma correlação fraca, mas estatisticamente significativa entre capacidade de memória de trabalho e pontuações no teste de recordação na segunda condição e nenhuma correlação entre capacidade de memória de trabalho e tempo de leitura. Análises exploratórias dos dados apontaram que a correlação observada foi influenciada pela proficiência de leitura em língua estrangeira e que a razão entre pontuações do teste de recordação e tempo de leitura é uma medida mais apropriada para a eficiência da compreensão leitora.

Accurate reading comprehension rate as an indicator of broad reading in students in first, second, and third grades.

The relationship between reading comprehension rate measures and broad reading skill development was examined using data from approximately 1425 students (grades 1-3). Students read 3 passages, from a pool of 30, and answered open-ended comprehension questions. Accurate reading comprehension rate (ARCR) was calculated by dividing the percentage of questions answered correctly (%QC) by seconds required to read the passage. Across all 30 passages, ARCR and its two components, %QC correct and time spent reading (1/seconds spent reading the passage), were significantly correlated with broad reading scores, with %QC resulting in the lowest correlations. Two sequential regressions supported previous findings which suggest that ARCR measures consistently produced meaningful incremental increases beyond %QC in the amount of variance explained in broad reading skill; however, ARCR produced small or no incremental increases beyond reading time. Discussion focuses on the importance of the measure of reading time embedded in brief accurate reading rate measures and directions for future research.

A Study on Anlaysis of Patch Test Positive Reactivity according to the Reading Time / 대한피부과학회지

Since the standard recommendation by the International Contact Dermatitis Research Group, day 2 and day 4 patch test readings have been performed by most dermatologists. However, nearly half of positive reactions can be observed only either on day 2 or on day 4 and the interpretation of these reactions is a problem. We compared the patch test reaction according to the reading time in 754 patients who had been performed patch test with standard battery and in 367 patients who had been performed patch test with various cosmetic allergens including cosmetics as is. The study result is summoized as follows : 1. Persistent reactions were observed in 59.8% patients tested with standard battery, however transient reactions were observed in 20.8% and delayed reaction in 19.4%. 2. Frequent allergens which showed transient reaction were carba mix, paraben mix, tego, imidazolidinyl urea, in order of frequency. Frequent allergens which showed late reactions were neomycin sulfate, quinidine mix, farmaldehyde, and imidazolidinyl urea. 3. Persistent reactions were observed in 47.1% patients. tested with cosmetics as is, however transient reactions were observed in 29.4% and delayed reactions in 23.5%, Facial skin care products showed the i,.st high positivity. 4. Persistent reactions were observed in 41.0% patients tested with cosmetics ingredients, however transient reactions were observed in 38.7% and delayed reactions in 23.0%. Frequent cosmetic allergens were cinnamic alcohol, cinnamic aldehyde, ylang ylang oil, sandawood oil, and benzyl salicylate. In conclusion, the reaction pattern in general was similar to the dats of others. However, the individual allergens showed transient reaction were somewhat different. Higher rate of transient reactions were observed in patch test positive reactions with cosmetic ingredients than standard battery and cosmetics as is.