ADR-Net: Context extraction network based on M-Net for medical image segmentation.

PURPOSE: Medical image segmentation is an essential component of medical image analysis. Accurate segmentation can assist doctors in diagnosis and relieve their fatigue. Although several image segmentation methods based on U-Net have been proposed, their performances have been observed to be suboptimal in the case of small-sized objects. To address this shortcoming, a novel network architecture is proposed in this study to enhance segmentation performance on small medical targets. METHODS: In this paper, we propose a joint multi-scale context attention network architecture to simultaneously capture higher level semantic information and spatial information. In order to obtain a greater number of feature maps during decoding, the network concatenates the images of side inputs by down-sampling during the encoding phase. In the bottleneck layer of the network, dense atrous convolution (DAC) and multi-scale residual pyramid pooling (RMP) modules are exploited to better capture high-level semantic information and spatial information. To improve the segmentation performance on small targets, the attention gate (AG) block is used to effectively suppress feature activation in uncorrelated regions and highlight the target area. RESULTS: The proposed model is first evaluated on the public dataset DRIVE, on which it performs significantly better than the basic framework in terms of sensitivity (SE), intersection-over-union (IOU), and area under the receiver operating characteristic curve (AUC). In particular, the SE and IOU are observed to increase by 7.46% and 5.97%, respectively. Further, the evaluation indices exhibit improvements compared to those of state-of-the-art methods as well, with SE and IOU increasing by 3.58% and 3.26%, respectively. Additionally, in order to demonstrate the generalizability of the proposed architecture, we evaluate our model on three other challenging datasets. The respective performances are observed to be better than those of state-of-the-art network architectures on the same datasets. Moreover, we use lung segmentation as a comparative experiment to demonstrate the transferability of the advantageous properties of the proposed approach in the context of small target segmentation to the segmentation of large targets. Finally, an ablation study is conducted to investigate the individual contributions of the AG block, the DAC block, and the RMP block to the performance of the network. CONCLUSIONS: The proposed method is evaluated on various datasets. Experimental results demonstrate that the proposed model performs better than state-of-the-art methods in medical image segmentation of small targets.

An extended research of crossmodal correspondence between color and sound in psychology and cognitive ergonomics.

Based on the existing research on sound symbolism and crossmodal correspondence, this study proposed an extended research on cross-modal correspondence between various sound attributes and color properties in a group of non-synesthetes. In Experiment 1, we assessed the associations between each property of sounds and colors. Twenty sounds with five auditory properties (pitch, roughness, sharpness, tempo and discontinuity), each varied in four levels, were used as the sound stimuli. Forty-nine colors with different hues, saturation and brightness were used to match to those sounds. Result revealed that besides pitch and tempo, roughness and sharpness also played roles in sound-color correspondence. Reaction times of sound-hue were a little longer than the reaction times of sound-lightness. In Experiment 2, a speeded target discrimination task was used to assess whether the associations between sound attributes and color properties could invoke natural cross-modal correspondence and improve participants' cognitive efficiency in cognitive tasks. Several typical sound-color pairings were selected according to the results of Experiment 1. Participants were divided into two groups (congruent and incongruent). In each trial participants had to judge whether the presented color could appropriately be associated with the sound stimuli. Result revealed that participants responded more quickly and accurately in the congruent group than in the incongruent group. It was also found that there was no significant difference in reaction times and error rates between sound-hue and sound-lightness. The results of Experiment 1 and 2 indicate the existence of a robust crossmodal correspondence between multiple attributes of sound and color, which also has strong influence on cognitive tasks. The inconsistency of the reaction times between sound-hue and sound-lightness in Experiment 1 and 2 is probably owing to the difference in experimental protocol, which indicates that the complexity of experiment design may be an important factor in crossmodal correspondence phenomena.

Time-Dependent Effects of Acute Exercise on University Students' Cognitive Performance in Temperate and Cold Environments.

Background: Few studies have examined the acute exercise-induced changes in cognitive performance in different thermal environments and the time course effects. Objective: Investigate the time-dependent effects of acute exercise on university students' processing speed, working memory and cognitive flexibility in temperate and cold environments. Method: Twenty male university students (age 23.5 ± 2.0 years) with moderate physical activity level participated in a repeated-measures within-subjects design. Processing speed, working memory and cognitive flexibility were assessed using CogState test battery at baseline (BASE), followed by a 45-min rest (REST), immediately after (EX) and 30 min after (POST-EX) 30-min moderate-intensity treadmill running in both temperate (TEMP; 25°C) and cold (COLD; 10°C) environments. Mean skin temperature (MST) and thermal sensation (TS) were also recorded. Two-way repeated measures ANOVA was performed to analyze each variable. Spearman's rho was used to identify the correlations between MST, TS and cognitive performance. Results: Reaction time (RT) of processing speed and working memory decreased immediately after exercise in both conditions (processing speed: p = 0.003; working memory: p = 0.007). The facilitating effects on processing speed disappeared within 30 min after exercise in TEMP (p = 0.163) and COLD (p = 0.667), while improvements on working memory remained 30 min after exercise in TEMP (p = 0.047), but not in COLD (p = 0.663). Though RT of cognitive flexibility reduced in both conditions (p = 0.003), no significance was found between EX and REST (p = 0.135). Increased MST and TS were significantly associated with reductions in processing speed RT (MST: r = -0.341, p < 0.001; TS: r = -0.262, p = 0.001) and working memory RT (MST: r = -0.282, p < 0.001; TS: r = -0.2229, p = 0.005), and improvements in working memory accuracy (MST: r = 0.249, p = 0.002; TS: r = 0.255, p = 0.001). Conclusion: The results demonstrate different time-dependent effects of acute exercise on cognition in TEMP and COLD. Our study reveals facilitating effects of exercise on university students' processing speed and working memory in both environments. However, in contrast to TEMP, effects on working memory in COLD are transient.

Household Fluorescent Lateral Flow Strip Platform for Sensitive and Quantitative Prognosis of Heart Failure Using Dual-Color Upconversion Nanoparticles.

Heart failure (HF) is the end-stage of cardiovascular diseases, which is associated with a high mortality rate and high readmission rate. Household early diagnosis and real-time prognosis of HF at bedside are of significant importance. Here, we developed a highly sensitive and quantitative household prognosis platform (termed as UC-LFS platform), integrating a smartphone-based reader with multiplexed upconversion fluorescent lateral flow strip (LFS). Dual-color core-shell upconversion nanoparticles (UCNPs) were synthesized as probes for simultaneously quantifying two target antigens associated with HF, i.e., brain natriuretic peptide (BNP) and suppression of tumorigenicity 2 (ST2). With the fluorescent LFS, we achieved the specific detection of BNP and ST2 antigens in spiked samples with detection limits of 5 pg/mL and 1 ng/mL, respectively, both of which are of one order lower than their clinical cutoff. Subsequently, a smartphone-based portable reader and an analysis app were developed, which could rapidly quantify the result and share prognosis results with doctors. To confirm the usage of UC-LFS platform for clinical samples, we detected 38 clinical serum samples using the platform and successfully detected the minimal concentration of 29.92 ng/mL for ST2 and 17.46 pg/mL for BNP in these clinical samples. Comparing the detection results from FDA approved clinical methods, we obtained a good linear correlation, indicating the practical reliability and stability of our developed UC-LFS platform. Therefore, the developed UC-LFS platform is demonstrated to be highly sensitive and specific for sample-to-answer prognosis of HF, which holds great potential for risk assessment and health monitoring of post-treatment patients at home.