Stable feature selection utilizing Graph Convolutional Neural Network and Layer-wise Relevance Propagation for biomarker discovery in breast cancer

High-throughput technologies are becoming increasingly important in discovering prognostic biomarkers and in identifying novel drug targets. With Mammaprint, Oncotype DX, and many other prognostic molecular signatures breast cancer is one of the paradigmatic examples of the utility of high-throughput data to deliver prognostic biomarkers, that can be represented in a form of a rather short gene list. Such gene lists can be obtained as a set of features (genes) that are important for the decisions of a Machine Learning (ML) method applied to high-dimensional gene expression data. Several studies have identified predictive gene lists for patient prognosis in breast cancer, but these lists are unstable and have only a few genes in common. Instability of feature selection impedes biological interpretability: genes that are relevant for cancer pathology should be members of any predictive gene list obtained for the same clinical type of patients. Stability and interpretability of selected features can be improved by including information on molecular networks in ML methods. Graph Convolutional Neural Network (GCNN) is a contemporary deep learning approach applicable to gene expression data structured by a prior knowledge molecular network. Layer-wise Relevance Propagation (LRP) and SHapley Additive exPlanations (SHAP) are methods to explain individual decisions of deep learning models. We used both GCNN+LRP and GCNN+SHAP techniques to construct feature sets by aggregating individual explanations. We suggest a methodology to systematically and quantitatively analyze the stability, the impact on the classification performance, and the interpretability of the selected feature sets. We used this methodology to compare GCNN+LRP to GCNN+SHAP and to more classical ML-based feature selection approaches. Utilizing a large breast cancer gene expression dataset we show that, while feature selection with SHAP is useful in applications where selected features have to be impactful for classification performance, among all studied methods GCNN+LRP delivers the most stable (reproducible) and interpretable gene lists.

TBACkp: HER2 expression status classification network focusing on intrinsic subenvironmental characteristics of breast cancer liver metastases

The HER2 expression status in breast cancer liver metastases is a crucial indicator for the diagnosis, treatment, and prognosis assessment of patients. And typical diagnosis involves assessing the HER2 expression status through invasive procedures like biopsy. However, this method has certain drawbacks, such as being difficult in obtaining tissue samples and requiring long examination periods. To address these limitations, we propose an AI-aided diagnostic model. This model enables rapid diagnosis. It diagnoses a patient's HER2 expression status on the basis of preprocessed images, which is the region of the lesion extracted from a CT image rather than from an actual tissue sample. The algorithm of the model adopts a parallel structure, including a Branch Block and a Trunk Block. The Branch Block is responsible for extracting the gradient characteristics between the tumor sub-environments, and the Trunk Block is for fusing the characteristics extracted by the Branch Block. The Branch Block contains CNN with self-attention, which combines the advantages of CNN and self-attention to extract more meticulous and comprehensive image features. And the Trunk Block is so designed that it fuses the extracted image feature information without affecting the transmission of the original image features. The Conv-Attention is used to calculate the attention in the Trunk Block, which uses kernel dot product and is responsible for providing the weight for the self-attention in the process of using convolution induced deviation calculation. Combined with the structure of the model and the method used, we refer to this model as TBACkp. The dataset comprises the enhanced abdominal CT images of 151 patients with liver metastases from breast cancer, together with the corresponding HER2 expression levels for each patient. The experimental results are as follows: (AUC: 0.915, ACC: 0.854, specificity: 0.809, precision: 0.863, recall: 0.881, F1-score: 0.872). The results demonstrate that this method can accurately assess the HER2 expression status in patients when compared with other advanced deep learning model.

Proceedings of the 2023 Viral Clearance Symposium, Session 2: Viral Clearance Strategy and Case Studies

This session deals with the rational design of viral clearance studies for biopharmaceuticals including recombinant proteins such as monoclonal antibodies and, as new in scope of the symposium, also viral clearance for adeno-associated viral (AAV) vectors. For recombinant proteins, large datasets were accumulated over the last decades and are intended to be used for accelerated product process development and streamlining of viral clearance studies. How to utilize prior knowledge in viral clearance validation and how it can be used in a risk assessment tool to decide whether additional virus clearance studies are necessary during product development is being addressed by three of the presentations of this session. This also includes an a priori intended design and generation of validation data for a new kind of detergent such as CG-110, to build up a platform dataset to be used as prior knowledge in future marketing application. Another presentation investigates the virus removal mechanism of a newly developed hydrophobic interaction chromatography (HIC) resin and demonstrates for highly hydrophobic antibodies appropriate reduction for a retrovirus and impurities in a defined process range in contrast to the moderate to poor virus reduction of recent HIC resins. The last two presentations deal with virus clearance approaches for AAV, which will become mandatory with approval of the ICH Q5A revision. Appropriate virus removal and virus inactivation procedures can be implemented into the manufacturing processes of AAV vectors including viral filtration, viral inactivation (e.g., heat inactivation), affinity chromatography, and anion-exchange chromatography with which it seems possible to achieve a good clearance for helper and also adventitious viruses. The heat treatment step can be even a robust step for adenovirus helper inactivation for AAV products when product characteristics and process conditions are understood.

HVS-Unsup: Unsupervised cervical cell instance segmentation method based on human visual simulation

Instance segmentation plays an important role in the automatic diagnosis of cervical cancer. Although deep learning-based instance segmentation methods can achieve outstanding performance, they need large amounts of labeled data. This results in a huge consumption of manpower and material resources. To solve this problem, we propose an unsupervised cervical cell instance segmentation method based on human visual simulation, named HVS-Unsup. Our method simulates the process of human cell recognition and incorporates prior knowledge of cervical cells. Specifically, firstly, we utilize prior knowledge to generate three types of pseudo labels for cervical cells. In this way, the unsupervised instance segmentation is transformed to a supervised task. Secondly, we design a Nucleus Enhanced Module (NEM) and a Mask-Assisted Segmentation module (MAS) to address problems of cell overlapping, adhesion, and even scenarios involving visually indistinguishable cases. NEM can accurately locate the nuclei by the nuclei attention feature maps generated by point-level pseudo labels, and MAS can reduce the interference from impurities by updating the weight of the shallow network through the dice loss. Next, we propose a Category-Wise droploss (CW-droploss) to reduce cell omissions in lower-contrast images. Finally, we employ an iterative self-training strategy to rectify mislabeled instances. Experimental results on our dataset MS-cellSeg, the public datasets Cx22 and ISBI2015 demonstrate that HVS-Unsup outperforms existing mainstream unsupervised cervical cell segmentation methods.

Predictions transform memories: How expected versus unexpected events are integrated or separated in memory

Our brains constantly generate predictions about the environment based on prior knowledge. Many of the events we experience are consistent with these predictions, while others might be inconsistent with prior knowledge and thus violate our predictions. To guide future behavior, the memory system must be able to strengthen, transform, or add to existing knowledge based on the accuracy of our predictions. We synthesize recent evidence suggesting that when an event is consistent with our predictions, it leads to neural integration between related memories, which is associated with enhanced associative memory, as well as memory biases. Prediction errors, in turn, can promote both neural integration and separation, and lead to multiple mnemonic outcomes. We review these findings and how they interact with factors such as memory reactivation, prediction error strength, and task goals, to offer insight into what determines memory for events that violate our predictions. In doing so, this review brings together recent neural and behavioral research to advance our understanding of how predictions shape memory, and why.

PKDN: Prior Knowledge Distillation Network for bronchoscopy diagnosis

Bronchoscopy plays a crucial role in diagnosing and treating lung diseases. The deep learning-based diagnostic system for bronchoscopic images can assist physicians in accurately and efficiently diagnosing lung diseases, enabling patients to undergo timely pathological examinations and receive appropriate treatment. However, the existing diagnostic methods overlook the utilization of prior knowledge of medical images, and the limited feature extraction capability hinders precise focus on lesion regions, consequently affecting the overall diagnostic effectiveness. To address these challenges, this paper proposes a prior knowledge distillation network (PKDN) for identifying lung diseases through bronchoscopic images. The proposed method extracts color and edge features from lesion images using the prior knowledge guidance module, and subsequently enhances spatial and channel features by employing the dynamic spatial attention module and gated channel attention module, respectively. Finally, the extracted features undergo refinement and self-regulation through feature distillation. Furthermore, decoupled distillation is implemented to balance the importance of target and non-target class distillation, thereby enhancing the diagnostic performance of the network. The effectiveness of the proposed method is validated on the bronchoscopic dataset provided by Harbin Medical University Cancer Hospital, which consists of 2,029 bronchoscopic images from 200 patients. Experimental results demonstrate that the proposed method achieves an accuracy of 94.78% and an AUC of 98.17%, outperforming other methods significantly in diagnostic performance. These results indicate that the computer-aided diagnostic system based on PKDN provides satisfactory accuracy in diagnosing lung diseases during bronchoscopy.

Prior knowledge about events depicted in scenes decreases oculomotor exploration

The visual input that the eyes receive usually contains temporally continuous information about unfolding events. Therefore, humans can accumulate knowledge about their current environment. Typical studies on scene perception, however, involve presenting multiple unrelated images and thereby render this accumulation unnecessary. Our study, instead, facilitated it and explored its effects. Specifically, we investigated how recently-accumulated prior knowledge affects gaze behavior. Participants viewed sequences of static film frames that contained several 'context frames' followed by a 'critical frame'. The context frames showed either events from which the situation depicted in the critical frame naturally followed, or events unrelated to this situation. Therefore, participants viewed identical critical frames while possessing prior knowledge that was either relevant or irrelevant to the frames' content. In the former case, participants' gaze behavior was slightly more exploratory, as revealed by seven gaze characteristics we analyzed. This result demonstrates that recently-gained prior knowledge reduces exploratory eye movements.

Aging attenuates the memory advantage for unexpected objects in real-world scenes

Across the adult lifespan memory processes are subject to pronounced changes. Prior knowledge and expectations might critically shape functional differences; however, corresponding findings have remained ambiguous so far. Here, we chose a tailored approach to scrutinize how schema (in-)congruencies affect older and younger adults' memory for objects embedded in real-world scenes, a scenario close to everyday life memory demands. A sample of 23 older (52-81 years) and 23 younger adults (18-38 years) freely viewed 60 photographs of scenes in which target objects were included that were either congruent or incongruent with the given context information. After a delay, recognition performance for those objects was determined. In addition, recognized objects had to be matched to the scene context in which they were previously presented. While we found schema violations beneficial for object recognition across age groups, the advantage was significantly less pronounced in older adults. We moreover observed an age-related congruency bias for matching objects to their original scene context. Our findings support a critical role of predictive processes for age-related memory differences and indicate enhanced weighting of predictions with age. We suggest that recent predictive processing theories provide a particularly useful framework to elaborate on age-related functional vulnerabilities as well as stability.

Source apportionment of fine particulate matter at a megacity in China, using an improved regularization supervised PMF model

The source apportionment of particulate matter plays an important role in solving the atmospheric particulate pollution. Positive matrix factorization (PMF) is a widely used source apportionment model. At present, high resolution online datasets are increasingly rich, but acquiring accurate and timely source apportionment results is still challenging. Integrating prior knowledge into modelling process is an effective solution and can yield reliable results. This study proposed an improved source apportionment method for the regularization supervised PMF model (RSPMF). This method leveraged actual source profile to guide factor profile for rapidly and automatically identifying source categories and quantifying source contributions. The results showed that the factor profile from RSPMF could be interpreted as seven factors and approach to actual source profile. Average source contributions were also an agreement between RSPMF and EPAPMF, including secondary nitrate (26 %, 27 %), secondary sulfate (23 %, 24 %), coal combustion (18 %, 18 %), vehicle exhaust (15 %, 15 %), biomass burning (10 %, 9 %), dust (5 %, 4 %), industrial emission (3 %, 3 %). The solutions of RSPMF also exhibited good generalizability during different episodes. This study reveals the superiority of supervised model, this model embeds prior knowledge into modelling process to guide model for obtaining more reliable results.

Lack of source memory as a potential marker of early assimilation of novel items into current knowledge

In daily life, humans process a plethora of new information that can be either consistent (familiar) or inconsistent (novel) with prior knowledge. Over time, both types of information can integrate into our accumulated knowledge base via distinct pathways. However, the mnemonic processes supporting the integration of information that is inconsistent with prior knowledge remain under-characterized. In the current study, we used functional magnetic resonance imaging (fMRI) to examine the initial assimilation of novel items into the semantic network. Participants saw three repetitions of adjective-noun word pairs that were either consistent or inconsistent with prior knowledge. Twenty-four hours later, they were presented with the same stimuli again while undergoing fMRI scans. Outside the scanner, participants completed a surprise recognition test. We found that when the episodic context associated with initially inconsistent items was irretrievable, the neural signature of these items was indistinguishable from that of consistent items. In contrast, initially inconsistent items with accessible episodic contexts showed neural signatures that differed from those associated with consistent items. We suggest that, at least one day post encoding, items inconsistent with prior knowledge can show early assimilation into the semantic network only when their episodic contexts become inaccessible during retrieval, thus evoking a sense of familiarity.

Does prior knowledge increase or decrease perceived visual complexity of texture images?

Previous research has shown that the perceived visual complexity of an image is correlated with understandability of the image. It was considered that prior knowledge of the contents of an image makes images easier to understand, and thus reduces perceived visual complexity. In the present study, we examined the effect of prior knowledge on perceived visual complexity of texture images. We designed an experiment in which participants observed and rated four texture images with different levels of complexity and understandability; one group of participants received prior knowledge in the form of verbal cues about the names of the target stimuli while the other group did not receive any information regarding image content. We found that the effect of prior knowledge on visual complexity perception varied for the different images. For an image with low initial complexity, if cued information about the image is three-dimensional or dynamic, prior knowledge does not decrease but instead increases the perceived visual complexity. Moreover, cues that increase perceived visual complexity can be verbal rather than visual cues.

Perceived satisfaction of university students with the use of chatbots as a tool for self-regulated learning

Chatbots are a promising resource for giving students feedback and helping them deploy metacognitive strategies in their learning processes. In this study we worked with a sample of 57 university students, 42 undergraduate and 15 Master's degree students in Health Sciences. A mixed research methodology was applied. The quantitative study analysed the influence of the variables educational level (undergraduate vs. master's degree) and level of prior knowledge on the frequency of chatbot use (low vs. average), learning outcomes, and satisfaction with the chatbot's usefulness. In addition, we examined whether the frequency of chatbot use depended on students' metacognitive strategies. The qualitative study analysed the students' suggestions for improvement to the chatbot and the type of questions it used. The results indicated that the level of degree being studied influenced the frequency of chatbot use and learning outcomes, with Master's students exhibiting higher levels of both, but levels of prior knowledge only influenced learning outcomes. Significant differences were also found in students' perceived satisfaction with the use of the chatbot, with Master's students scoring higher, but not with respect to the level of prior knowledge. No conclusive results were found regarding frequency of chatbot use and the levels of students' metacognitive strategies. Further studies are needed to guide this research based on the students' suggestions for improvement.

Temporal-based Swin Transformer network for workflow recognition of surgical video

PURPOSE

Surgical workflow recognition has emerged as an important part of computer-assisted intervention systems for the modern operating room, which also is a very challenging problem. Although the CNN-based approach achieves excellent performance, it does not learn global and long-range semantic information interactions well due to the inductive bias inherent in convolution.

METHODS

In this paper, we propose a temporal-based Swin Transformer network (TSTNet) for the surgical video workflow recognition task. TSTNet contains two main parts: the Swin Transformer and the LSTM. The Swin Transformer incorporates the attention mechanism to encode remote dependencies and learn highly expressive representations. The LSTM is capable of learning long-range dependencies and is used to extract temporal information. The TSTNet organically combines the two components to extract spatiotemporal features that contain more contextual information. In particular, based on a full understanding of the natural features of the surgical video, we propose a priori revision algorithm (PRA) using a priori information about the sequence of the surgical phase. This strategy optimizes the output of TSTNet and further improves the recognition performance.

RESULTS

We conduct extensive experiments using the Cholec80 dataset to validate the effectiveness of the TSTNet-PRA method. Our method achieves excellent performance on the Cholec80 dataset, which accuracy is up to 92.8% and greatly exceeds the state-of-the-art methods.

CONCLUSION

By modelling remote temporal information and multi-scale visual information, we propose the TSTNet-PRA method. It was evaluated on a large public dataset, showing a high recognition capability superior to other spatiotemporal networks.

Prior Knowledge Guided Ultra-high Dimensional Variable Screening with Application to Neuroimaging Data

Variable screening is a powerful and efficient tool for dimension reduction under ultrahigh dimensional settings. However, most existing methods overlook useful prior knowledge in specific applications. In this work, from a Bayesian modeling perspective, we develop a unified variable screening procedure for the linear regression model. We discuss different constructions of posterior mean screening (PMS) statistics to incorporate different types of prior knowledge according to specific applications. With non-informative prior specifications, PMS is equivalent to high-dimensional ordinary least-square projections (HOLP). We establish the screening consistency property for PMS with different types of prior knowledge. We show that PMS is robust to prior misspecifications; and when the prior knowledge provides correct information on summarizing the true parameter settings, PMS can substantially improve the selection accuracy compared to HOLP and other existing methods. We illustrate our method with extensive simulation studies and an analysis of neuroimaging data.

What affects gestural learning in children with and without Autism? The role of prior knowledge and imitation

The present study examined whether prior knowledge to the learning target and imitation during learning affected learning outcomes in preschool children with autism spectrum disorder (ASD, N = 22) compared to their typically developing (TD, N = 15) peers. Children's gestural skills in recognizing and producing the target gestures before and after the training, as well as their imitative behavior during the training were coded. Results showed that consistent prior knowledge benefited gestural learning in both groups. Besides, only children with ASD were hindered by inconsistent prior knowledge. Notably, the effect of imitation was not significant in the ASD group. In conclusion, the learning process in children with ASD may differ from those with typical development, suggesting special-designed interventions are required.

Effects of limited previously acquired information about falling height on lower limb biomechanics when individuals are landing with limited visual input

BACKGROUND

Inhibitions in the acquisition of accurate information about the environment can affect control of the lower extremities and lead to anterior cruciate ligament injury. This study aimed to clarify the effects of limited prior knowledge of the height of the fall, as well as limited visual input, on lower limb and trunk motion and ground reaction force during landing.

METHODS

Twenty healthy university students were recruited. Drop landings from a 30-cm platform were measured under three conditions: (1) unknown, without prior knowledge of the height of the fall and without visual input; (2) known, with prior knowledge of the height of the fall and without visual input; and (3) control, with prior knowledge of the height of the fall and visual input.

FINDINGS

In the unknown condition, the peak ground reaction force for the vertical and posterior directions was significantly higher than that in the known and control conditions; leg and knee stiffness, ankle joint work, and joint flexion motion of the knee, ankle, and trunk after landing were decreased as well. In the known condition, there were no significant differences in leg and knee stiffness and vertical ground reaction force compared to the control condition.

INTERPRETATION

The results of this study indicate that the risk of anterior cruciate ligament injury during landing increases when individuals have limited visual input and prior knowledge of the height of the fall. This finding suggests that an accurate perception of the surrounding environment may help prevent anterior cruciate ligament injuries.

Emotion schema effects on associative memory differ across emotion categories at the behavioural, physiological and neural level: Emotion schema effects on associative memory differs for disgust and fear

Previous behavioural and neuroimaging studies have consistently reported that memory is enhanced for associations congruent or incongruent with the structure of prior knowledge, termed as schemas. However, it remains unclear if similar effects arise with emotion-related associations, and whether they depend on the type of emotions. Here, we addressed this question using a novel face-word pair association paradigm combined with fMRI and eye-tracking techniques. In two independent studies, we demonstrated and replicated that both congruency with emotion schemas and emotion category interact to affect associative memory. Overall, memory retrieval was higher for faces from pairs congruent vs. incongruent with emotion schemas, paralleled by a greater recruitment of left inferior frontal gyrus (IFG) during successful encoding. However, emotion schema effects differed across two negative emotion categories. Disgust was remembered better than fear, and only disgust activated left IFG stronger during encoding of congruent vs. incongruent pairs, suggestive of deeper semantic processing for the associations. On the contrary, encoding of congruent fear vs. disgust-related pairs was accompanied with greater activity in right fusiform gyrus (FG), suggesting a stronger sensory processing of faces. In addition, successful memory formation for congruent disgust pairs was associated with a higher pupil dilation index related to sympathetic activation, longer gaze time on words compared to faces, and more gaze switches between paired words and faces. This was reversed for fear-related congruent pairs where the faces attracted longer gaze time (compared to words). Overall, our results provide converging evidence from behavioural, physiological, and neural measures to suggest that congruency with available emotion schemas influence memory associations in a similar manner to semantic schemas. However, these effects vary across distinct emotion categories, pointing to a differential role of semantic processing and visual attention processes in the modulation of memory by disgust and fear, respectively.

When the "Tabula" is Anything but "Rasa:" What Determines Performance in the Auditory Statistical Learning Task?

How does prior linguistic knowledge modulate learning in verbal auditory statistical learning (SL) tasks? Here, we address this question by assessing to what extent the frequency of syllabic co‐occurrences in the learners’ native language determines SL performance. We computed the frequency of co‐occurrences of syllables in spoken Spanish through a transliterated corpus, and used this measure to construct two artificial familiarization streams. One stream was constructed by embedding pseudowords with high co‐occurrence frequency in Spanish (“Spanish‐like” condition), the other by embedding pseudowords with low co‐occurrence frequency (“Spanish‐unlike” condition). Native Spanish‐speaking participants listened to one of the two streams, and were tested in an old/new identification task to examine their ability to discriminate the embedded pseudowords from foils. Our results show that performance in the verbal auditory SL (ASL) task was significantly influenced by the frequency of syllabic co‐occurrences in Spanish: When the embedded pseudowords were more “Spanish‐like,” participants were better able to identify them as part of the stream. These findings demonstrate that learners’ task performance in verbal ASL tasks changes as a function of the artificial language's similarity to their native language, and highlight how linguistic prior knowledge biases the learning of regularities.

Parameter Optimization for Ion Channel Models: Integrating New Data with Known Channel Properties

Ion channels play a central role in membrane physiology, but to fully understand how they operate, one must have accurate kinetic mechanisms. Estimating kinetics is not trivial when the mechanism is complex, and a large number of parameters must be extracted from data. Furthermore, the information contained in the data is often limited, and the model may not be fully determined. The solution is to reduce the number of parameters and to estimate them in such a way that they not only describe well the new data but also agree with the existing knowledge. In a previous study, we presented a comprehensive formalism for estimating kinetic parameters subject to a variety of explicit and implicit constraints that define quantitative relationships between parameters and describe specific mechanism properties. Here, we introduce the reader to the QuB software, which implements this constraining formalism. QuB features a powerful visual interface and a high-level scripting language that can be used to formulate kinetic models and constraints of arbitrary complexity, and to efficiently estimate the parameters from a variety of experimental data.

When perception is stronger than physics: Perceptual similarities rather than laws of physics govern the perception of interacting objects

When several multistable displays are viewed simultaneously, their perception is synchronized, as they tend to be in the same perceptual state. Here, we investigated the possibility that perception may reflect embedded statistical knowledge of physical interaction between objects for specific combinations of displays and layouts. We used a novel display with two ambiguously rotating gears and an ambiguous walker-on-a-ball display. Both stimuli produce a physically congruent perception when an interaction is possible (i.e., gears counterrotate, and the ball rolls under the walker's feet). Next, we gradually manipulated the stimuli to either introduce abrupt changes to the potential physical interaction between objects or keep it constant despite changes in the visual stimulus. We characterized the data using four different models that assumed (1) independence of perception of the stimulus, (2) dependence on the stimulus's properties, (3) dependence on physical configuration alone, and (4) an interaction between stimulus properties and a physical configuration. We observed that for the ambiguous gears, the perception was correlated with the stimulus changes rather than with the possibility of physical interaction. The perception of walker-on-a-ball was independent of the stimulus but depended instead on whether participants responded about a relative motion of two objects (perception was biased towards physically congruent motion) or the absolute motion of the walker alone (perception was independent of the rotation of the ball). None of the two experiments supported the idea of embedded knowledge of physical interaction.

Fine-tuning deep learning model parameters for improved super-resolution of dynamic MRI with prior-knowledge

Dynamic imaging is a beneficial tool for interventions to assess physiological changes. Nonetheless during dynamic MRI, while achieving a high temporal resolution, the spatial resolution is compromised. To overcome this spatio-temporal trade-off, this research presents a super-resolution (SR) MRI reconstruction with prior knowledge based fine-tuning to maximise spatial information while reducing the required scan-time for dynamic MRIs. A U-Net based network with perceptual loss is trained on a benchmark dataset and fine-tuned using one subject-specific static high resolution MRI as prior knowledge to obtain high resolution dynamic images during the inference stage. 3D dynamic data for three subjects were acquired with different parameters to test the generalisation capabilities of the network. The method was tested for different levels of in-plane undersampling for dynamic MRI. The reconstructed dynamic SR results after fine-tuning showed higher similarity with the high resolution ground-truth, while quantitatively achieving statistically significant improvement. The average SSIM of the lowest resolution experimented during this research (6.25% of the k-space) before and after fine-tuning were 0.939 ± 0.008 and 0.957 ± 0.006 respectively. This could theoretically result in an acceleration factor of 16, which can potentially be acquired in less than half a second. The proposed approach shows that the super-resolution MRI reconstruction with prior-information can alleviate the spatio-temporal trade-off in dynamic MRI, even for high acceleration factors.

Building memories on prior knowledge: behavioral and fMRI evidence of impairment in early Alzheimer's disease

Impaired memory is a hallmark of prodromal Alzheimer's disease (AD). Prior knowledge associated with the memoranda improves memory in healthy individuals, but we ignore whether the same occurs in early AD. We used functional MRI to investigate whether prior knowledge enhances memory encoding in early AD, and whether the nature of this prior knowledge matters. Patients with early AD and Controls underwent a task-based fMRI experiment where they learned face-scene associations. Famous faces carried pre-experimental knowledge (PEK), while unknown faces with which participants were familiarized prior to learning carried experimental knowledge (EK). Surprisingly, PEK strongly enhanced subsequent memory in healthy controls, but importantly not in patients. Partly nonoverlapping brain networks supported PEK vs. EK associative encoding in healthy controls. No such networks were identified in patients. In addition, patients displayed impaired activation in a right sub hippocampal region where activity predicted successful associative memory formation for PEK stimuli. Despite the limited sample sizes of this study, these findings suggest that the role prior knowledge in new learning might have been so far overlooked and underestimated in AD patients. Prior knowledge may drive critical differences in the way healthy elderly and early AD patients learn novel associations.

Examining the transition of novel information toward familiarity

Throughout their lives, humans encounter multiple instances of new information that can be inconsistent with prior knowledge (novel). Over time, the once-novel information becomes integrated into their established knowledge base, shifting from novelty to familiarity. In this study, we investigated the processes by which the first steps of this transition take place. We hypothesized that the neural representations of initially novel items gradually change over the course of repeated presentations, expressing a shift toward familiarity. We further assumed that this shift could be traced by examining neural patterns using fMRI. In two experiments, while being scanned, participants read noun-adjective word pairs that were either consistent or inconsistent with their prior knowledge. Stimuli were repeated 3-6 times within the scans. Employing mass univariate and multivariate similarity analyses, we showed that the neural representations associated with the initial presentation of familiar versus novel objects differed in lateral frontal and temporal regions, the medial prefrontal cortex, and the medial temporal lobe. Importantly, the neural representations of novel stimuli gradually changed throughout repetitions until they became indistinguishable from their respective familiar items. We interpret these findings as indicating that an early phase of familiarization can be completed within a few repetitions. This initial familiarization can then serve as the prerequisite to the integration of novel items into existing knowledge. Future empirical and theoretical works can build on the current findings to develop a comprehensive model of the transition from novelty to familiarity.

Specificity effects in reasoning with counterintuitive and arbitrary conditionals

When people have prior knowledge about an inference, they accept conclusions from specific conditionals (e.g., “If Jack does sports, then Jack loses weight”) more strongly than for unspecific conditionals (e.g., “If a person does sports, then the person loses weight”). But can specific phrasings also elevate the acceptance of conclusions from unbelievable conditionals? In Experiment 1, we varied the specificity of counterintuitive conditionals, which described the opposite of what is expected according to everyday experiences (“If Lena/a person studies hard, then Lena/the person will not do well on the test”). In Experiment 2, we varied the specificity of arbitrary conditionals, which had no obvious link between antecedent and consequent (“If Mary/a person goes shopping, then Mary/ the person gets pimples”). All conditionals were embedded in MP and AC inferences. Participants were instructed to reason as in daily life and to evaluate the conclusions on a 7-point Likert scale. Our results showed a specificity effect in both experiments: participants gave higher acceptance ratings for specific than for unspecific conditionals.

Prior episodic learning and the efficacy of retrieval practice

In three experiments we investigated how the level of study-based, episodic knowledge influences the efficacy of subsequent retrieval practice (testing) as a learning event. Possibilities are that the efficacy of a test, relative to a restudy control, decreases, increases, or is independent of the degree of prior study-based learning. The degree of study-based learning was manipulated by varying the number of item repetitions in the initial study phase between one and eight. Predictions of the dual-memory model of test-enhanced learning for the case of one study-phase repetition were used as a reference. Results support the hypothesis that the advantage of testing over restudy is independent of the degree of prior episodic learning, and they suggest that educators can apply cued-recall testing with the expectation that its efficacy is similar across varying levels of prior content learning. Implications for testing effect theory are discussed.

Comprior: facilitating the implementation and automated benchmarking of prior knowledge-based feature selection approaches on gene expression data sets

BACKGROUND

Reproducible benchmarking is important for assessing the effectiveness of novel feature selection approaches applied on gene expression data, especially for prior knowledge approaches that incorporate biological information from online knowledge bases. However, no full-fledged benchmarking system exists that is extensible, provides built-in feature selection approaches, and a comprehensive result assessment encompassing classification performance, robustness, and biological relevance. Moreover, the particular needs of prior knowledge feature selection approaches, i.e. uniform access to knowledge bases, are not addressed. As a consequence, prior knowledge approaches are not evaluated amongst each other, leaving open questions regarding their effectiveness.

RESULTS

We present the Comprior benchmark tool, which facilitates the rapid development and effortless benchmarking of feature selection approaches, with a special focus on prior knowledge approaches. Comprior is extensible by custom approaches, offers built-in standard feature selection approaches, enables uniform access to multiple knowledge bases, and provides a customizable evaluation infrastructure to compare multiple feature selection approaches regarding their classification performance, robustness, runtime, and biological relevance.

CONCLUSION

Comprior allows reproducible benchmarking especially of prior knowledge approaches, which facilitates their applicability and for the first time enables a comprehensive assessment of their effectiveness.

Protein analysis and stability: Overcoming trial-and-error by grouping according to physicochemical properties

Today proteins are possibly the most important class of substances. Yet new tasks for proteins are still often solved by trial-and-error approaches. However, in some areas these euphemistically called "screening approaches" are not suitable. E.g. stability tests just take too long and therefore require a more strategic, target-orientated concept. This concept is available by grouping proteins according to their physicochemical properties and then pulling out the right drawer for new tasks. These properties include size, then charge and hydrophobicity as well as their patchinesses, and the degree of order. In addition, solubility, the content of (free) enthalpy, aromatic-amino-acid- and α/β-frequency as well as helix capping, and corresponding patchiness, the number of specific motifs and domains as well as the typical concentration range can be helpful to discriminate between different groups of proteins. Analyzing correlations will reduce the necessary amount of parameters and additional ones, which may be still undiscovered at the present time, can be identified looking at protein subgroups with similar physicochemical properties which still behave heterogeneously. Step-by-step the methodology will be improved. Possibly protein stability will be the driver of this process, but all other areas such as production, purification and analytics including sample pre-treatment and the choice of appropriate separation conditions for e.g. chromatography and electrophoresis will profit from a rational strategy.

Why are the batteries in the microwave?: Use of semantic information under uncertainty in a search task

A major problem in human cognition is to understand how newly acquired information and long-standing beliefs about the environment combine to make decisions and plan behaviors. Over-dependence on long-standing beliefs may be a significant source of suboptimal decision-making in unusual circumstances. While the contribution of long-standing beliefs about the environment to search in real-world scenes is well-studied, less is known about how new evidence informs search decisions, and it is unclear whether the two sources of information are used together optimally to guide search. The present study expanded on the literature on semantic guidance in visual search by modeling a Bayesian ideal observer's use of long-standing semantic beliefs and recent experience in an active search task. The ability to adjust expectations to the task environment was simulated using the Bayesian ideal observer, and subjects' performance was compared to ideal observers that depended on prior knowledge and recent experience to varying degrees. Target locations were either congruent with scene semantics, incongruent with what would be expected from scene semantics, or random. Half of the subjects were able to learn to search for the target in incongruent locations over repeated experimental sessions when it was optimal to do so. These results suggest that searchers can learn to prioritize recent experience over knowledge of scenes in a near-optimal fashion when it is beneficial to do so, as long as the evidence from recent experience was learnable.

Explaining decisions of graph convolutional neural networks: patient-specific molecular subnetworks responsible for metastasis prediction in breast cancer

Background

Contemporary deep learning approaches show cutting-edge performance in a variety of complex prediction tasks. Nonetheless, the application of deep learning in healthcare remains limited since deep learning methods are often considered as non-interpretable black-box models. However, the machine learning community made recent elaborations on interpretability methods explaining data point-specific decisions of deep learning techniques. We believe that such explanations can assist the need in personalized precision medicine decisions via explaining patient-specific predictions.

Methods

Layer-wise Relevance Propagation (LRP) is a technique to explain decisions of deep learning methods. It is widely used to interpret Convolutional Neural Networks (CNNs) applied on image data. Recently, CNNs started to extend towards non-Euclidean domains like graphs. Molecular networks are commonly represented as graphs detailing interactions between molecules. Gene expression data can be assigned to the vertices of these graphs. In other words, gene expression data can be structured by utilizing molecular network information as prior knowledge. Graph-CNNs can be applied to structured gene expression data, for example, to predict metastatic events in breast cancer. Therefore, there is a need for explanations showing which part of a molecular network is relevant for predicting an event, e.g., distant metastasis in cancer, for each individual patient.

Results

We extended the procedure of LRP to make it available for Graph-CNN and tested its applicability on a large breast cancer dataset. We present Graph Layer-wise Relevance Propagation (GLRP) as a new method to explain the decisions made by Graph-CNNs. We demonstrate a sanity check of the developed GLRP on a hand-written digits dataset and then apply the method on gene expression data. We show that GLRP provides patient-specific molecular subnetworks that largely agree with clinical knowledge and identify common as well as novel, and potentially druggable, drivers of tumor progression.

Conclusions

The developed method could be potentially highly useful on interpreting classification results in the context of different omics data and prior knowledge molecular networks on the individual patient level, as for example in precision medicine approaches or a molecular tumor board.

Supplementary Information

The online version contains supplementary material available at (10.1186/s13073-021-00845-7).

Individual differences in executive function and learning: The role of knowledge type and conflict with prior knowledge

Executive function (EF) predicts children's academic achievement; however, less is known about the relation between EF and the actual learning process. The current study examined how aspects of the material to be learned-the type of information and the amount of conflict between the content to be learned and children's prior knowledge-influence the relation between individual differences in EF and learning. Typically developing 4-year-olds (N = 61) completed a battery of EF tasks and several animal learning tasks that varied on the type of information being learned (factual vs. conceptual) and the amount of conflict with the learners' prior knowledge (no prior knowledge vs. no conflicting prior knowledge vs. conflicting prior knowledge). Individual differences in EF predicted children's overall learning, controlling for age, verbal IQ, and prior knowledge. Children's working memory and cognitive flexibility skills predicted their conceptual learning, whereas children's inhibitory control skills predicted their factual learning. In addition, individual differences in EF mattered more for children's learning of information that conflicted with their prior knowledge. These findings suggest that there may be differential relations between EF and learning depending on whether factual or conceptual information is being taught and the degree of conceptual change that is required. A better understanding of these different relations serves as an essential foundation for future research designed to create more effective academic interventions to optimize children's learning.

Brain activity predicts future learning success in intensive second language listening training

This study explores neural mechanisms underlying how prior knowledge gained from pre-listening transcript reading helps comprehend fast-rate speech in a second language (L2) and applies to L2 learning. Top-down predictive processing by prior knowledge may play an important role in L2 speech comprehension and improving listening skill. By manipulating the pre-listening transcript effect (pre-listening transcript reading [TR] vs. no transcript reading [NTR]) and type of languages (first language (L1) vs. L2), we measured brain activity in L2 learners, who performed fast-rate listening comprehension tasks during functional magnetic resonance imaging. Thereafter, we examined whether TR_L2-specific brain activity can predict individual learning success after an intensive listening training. The left angular and superior temporal gyri were key areas responsible for integrating prior knowledge to sensory input. Activity in these areas correlated significantly with gain scores on subsequent training, indicating that brain activity related to prior knowledge-sensory input integration predicts future learning success.

Dynamic parameter estimation and prediction over consecutive scales, based on moving horizon estimation: applied to an industrial cell culture seed train

Bioprocess modeling has become a useful tool for prediction of the process future with the aim to deduce operating decisions (e.g. transfer or feeds). Due to variabilities, which often occur between and within batches, updating (re-estimation) of model parameters is required at certain time intervals (dynamic parameter estimation) to obtain reliable predictions. This can be challenging in the presence of low sampling frequencies (e.g. every 24 h), different consecutive scales and large measurement errors, as in the case of cell culture seed trains. This contribution presents an iterative learning workflow which generates and incorporates knowledge concerning cell growth during the process by using a moving horizon estimation (MHE) approach for updating of model parameters. This estimation technique is compared to a classical weighted least squares estimation (WLSE) approach in the context of model updating over three consecutive cultivation scales (40–2160 L) of an industrial cell culture seed train. Both techniques were investigated regarding robustness concerning the aforementioned challenges and the required amount of experimental data (estimation horizon). It is shown how the proposed MHE can deal with the aforementioned difficulties by the integration of prior knowledge, even if only data at two sampling points are available, outperforming the classical WLSE approach. This workflow allows to adequately integrate current process behavior into the model and can therefore be a suitable component of a digital twin.

DIKA-Nets: Domain-invariant knowledge-guided attention networks for brain skull stripping of early developing macaques

As non-human primates, macaques have a close phylogenetic relationship to human beings and have been proven to be a valuable and widely used animal model in human neuroscience research. Accurate skull stripping (aka. brain extraction) of brain magnetic resonance imaging (MRI) is a crucial prerequisite in neuroimaging analysis of macaques. Most of the current skull stripping methods can achieve satisfactory results for human brains, but when applied to macaque brains, especially during early brain development, the results are often unsatisfactory. In fact, the early dynamic, regionally-heterogeneous development of macaque brains, accompanied by poor and age-related contrast between different anatomical structures, poses significant challenges for accurate skull stripping. To overcome these challenges, we propose a fully-automated framework to effectively fuse the age-specific intensity information and domain-invariant prior knowledge as important guiding information for robust skull stripping of developing macaques from 0 to 36 months of age. Specifically, we generate Signed Distance Map (SDM) and Center of Gravity Distance Map (CGDM) based on the intermediate segmentation results as guidance. Instead of using local convolution, we fuse all information using the Dual Self-Attention Module (DSAM), which can capture global spatial and channel-dependent information of feature maps. To extensively evaluate the performance, we adopt two relatively-large challenging MRI datasets from rhesus macaques and cynomolgus macaques, respectively, with a total of 361 scans from two different scanners with different imaging protocols. We perform cross-validation by using one dataset for training and the other one for testing. Our method outperforms five popular brain extraction tools and three deep-learning-based methods on cross-source MRI datasets without any transfer learning.

Compressed sensing and the use of phased array coils in 23Na MRI: a comparison of a SENSE-based and an individually combined multi-channel reconstruction

PURPOSE

To implement and to evaluate a compressed sensing (CS) reconstruction algorithm based on the sensitivity encoding (SENSE) combination scheme (CS-SENSE), used to reconstruct sodium magnetic resonance imaging (²³Na MRI) multi-channel breast data sets.

METHODS

In a simulation study, the CS-SENSE algorithm was tested and optimized by evaluating the structural similarity (SSIM) and the normalized root-mean-square error (NRMSE) for different regularizations and different undersampling factors (USF=1.8/3.6/7.2/14.4). Subsequently, the algorithm was applied to data from in vivo measurements of the healthy female breast (n=3) acquired at 7T. Moreover, the proposed CS-SENSE algorithm was compared to a previously published CS algorithm (CS-IND).

RESULTS

The CS-SENSE reconstruction leads to an increased image quality for all undersampling factors and employed regularizations. Especially if a simple 2^nd order total variation is chosen as sparsity transformation, the CS-SENSE reconstruction increases the image quality of highly undersampled data sets (CS-SENSE: SSIM_USF=7.2=0.234, NRMSE_USF=7.2=0.491 vs. CS-IND: SSIM_USF=7.2=0.201, NRMSE_USF=7.2=0.506).

CONCLUSION

The CS-SENSE reconstruction supersedes the need of CS weighting factors for each channel as well as a method to combine single channel data. The CS-SENSE algorithm can be used to reconstruct undersampled data sets with increased image quality. This can be exploited to reduce total acquisition times in ²³Na MRI.

Recommendations for the analysis of gene expression data to identify intrinsic differences between similar tissues

Identifying genes involved in functional differences between similar tissues from expression profiles is challenging, because the expected differences in expression levels are small. To exemplify this challenge, we studied the expression profiles of two skeletal muscles, deltoid and biceps, in healthy individuals. We provide a series of guides and recommendations for the analysis of this type of studies. These include how to account for batch effects and inter-individual differences to optimize the detection of gene signatures associated with tissue function. We provide guidance on the selection of optimal settings for constructing gene co-expression networks through parameter sweeps of settings and calculation of the overlap with an established knowledge network. Our main recommendation is to use a combination of the data-driven approaches, such as differential gene expression analysis and gene co-expression network analysis, and hypothesis-driven approaches, such as gene set connectivity analysis. Accordingly, we detected differences in metabolic gene expression between deltoid and biceps that were supported by both data- and hypothesis-driven approaches. Finally, we provide a bioinformatic framework that support the biological interpretation of expression profiles from related tissues from this combination of approaches, which is available at github.com/tabbassidaloii/AnalysisFrameworkSimilarTissues.

Influencing public health policy with data-informed mathematical models of infectious diseases: Recent developments and new challenges

Modern data and computational resources, coupled with algorithmic and theoretical advances to exploit these, allow disease dynamic models to be parameterised with increasing detail and accuracy. While this enhances models' usefulness in prediction and policy, major challenges remain. In particular, lack of identifiability of a model's parameters may limit the usefulness of the model. While lack of parameter identifiability may be resolved through incorporation into an inference procedure of prior knowledge, formulating such knowledge is often difficult. Furthermore, there are practical challenges associated with acquiring data of sufficient quantity and quality. Here, we discuss recent progress on these issues.

Super-resolved water/fat image reconstruction based on single-shot spatiotemporally encoded MRI

Single-shot spatiotemporally encoded (SPEN) MRI has been validated to possess considerable performance in both spatial and temporal resolution. Water/fat separation is essential for MRI applications in which only water signal is needed. In this article, a super-resolved water/fat image reconstruction method (dubbed SWAF) combined prior knowledge was developed based on single-shot SPEN MRI. The point spread function of spatiotemporal encoding under multiple chemical shifts situation was derived and used for constructing an equation for SWAF image reconstruction. By processing the prior chemical shift information with filtering operation, an initial spin density profile of water/fat and a weighting matrix for water/fat residual artifacts suppression were obtained to guide the reconstruction process. A l₁ norm minimization problem with regularization was exploited to reconstruct separated water/fat images with high spatial resolution and less residual/aliasing artifacts. Numeric simulation and experiments on water-oil phantom and rat abdomen/neck imaging demonstrated the effectiveness and robustness of this new method. The SWAF method proposed herein would promote the application of SPEN MRI in the cases where water/fat separation is required.

Effect of dynamic visual motion on perception of postural vertical through the modulation of prior knowledge of gravity

To internally estimate gravitational direction and body orientation, the central nervous system considers several sensory inputs from the periphery and prior knowledge of gravity. It is hypothesized that the modulation of visual inputs, supplying indirect information of gravity, affects the prior knowledge established internally by other sensory inputs from vestibular and somatosensory systems, leading to the alteration of perceived body orientation relative to gravity. In order to test the hypothesis, we examined the effect of presenting a visual motion stimulus during a whole-body static tilt on the subsequent evaluation of the perceived postural vertical. Fifteen subjects watched a target moving along the body longitudinal axis directing from head to feet with constant downward acceleration (CA condition) or constant velocity (CV condition), or they did not receive any visual stimulation (NV condition) during the whole-body static tilt. Subsequently, the direction of the subjective postural vertical (SPV) was evaluated. The result showed that the SPV in the CA condition was significantly tilted toward the direction of the preceding tilt compared to that in the NV condition while those in the CV and NV conditions were not significantly different. The present result suggests that dynamic visual motion along body longitudinal axis with downward acceleration can modulate prior knowledge of gravity, and in turn this affects the perception of body verticality.

Improving neural protein-protein interaction extraction with knowledge selection

Protein-protein interaction (PPI) extraction from published scientific literature provides additional support for precision medicine efforts. Meanwhile, knowledge bases (KBs) contain huge amounts of structured information of protein entities and their relations, which can be encoded in entity and relation embeddings to help PPI extraction. However, the prior knowledge of protein-protein pairs must be selectively used so that it is suitable for different contexts. This paper proposes a Knowledge Selection Model (KSM) to fuse the selected prior knowledge and context information for PPI extraction. Firstly, two Transformers encode the context sequence of a protein pair according to each protein embedding, respectively. Then, the two outputs are fed to a mutual attention to capture the important context features towards the protein pair. Next, the context features are used to distill the relation embedding by a knowledge selector. Finally, the selected relation embedding and the context features are concatenated for PPI extraction. Experiments on the BioCreative VI PPI dataset show that KSM achieves a new state-of-the-art performance (38.08 % F1-score) by adding knowledge selection.

Incorporating prior shape knowledge via data-driven loss model to improve 3D liver segmentation in deep CNNs

PURPOSE

Convolutional neural networks (CNNs) have obtained enormous success in liver segmentation. However, there are several challenges, including low-contrast images, and large variations in the shape, and appearance of the liver. Incorporating prior knowledge in deep CNN models improves their performance and generalization.

METHODS

A convolutional denoising auto-encoder is utilized to learn global information about 3D liver shapes in a low-dimensional latent space. Then, the deep data-driven knowledge is used to define a loss function and combine it with the Dice loss in the main segmentation model. The resultant hybrid model would be forced to learn the global shape information as prior knowledge, while it tries to produce accurate results and increase the Dice score.

RESULTS

The proposed training strategy improved the performance of the 3D U-Net model and reached the Dice score of 97.62% on the Sliver07-I liver dataset, which is competitive to the state-of-the-art automatic segmentation methods. The proposed algorithm enhanced the generalization and robustness of the hybrid model and outperformed the 3D U-Net model in the prediction of unseen images.

CONCLUSIONS

The results indicate that the incorporation of prior shape knowledge enhances liver segmentation tasks in deep CNN models. The proposed method improves the generalization and robustness of the hybrid model due to the abstract features provided by the data-driven loss model.

Determining the impact of pre-lecture educational video on comprehension of a difficult gross anatomy lecture

Objective

Students commonly perceive gross anatomy lectures as difficult because they contain complex information that requires three-dimensional visualisation in order to be understood. Without prior preparation, a gross anatomy topic expounded via lecture can be cognitively challenging. Hence, this study aimed to investigate the impact of a pre-lecture activity in the form of viewing a video on students' lecture comprehension.

Method

A quasi-experimental study was conducted using 254 first-year medical students with no prior exposure to the lecture topic during the 2016/17 and 2017/18 academic sessions. The students from each batch were divided into two groups and exposed to different video material. Group A watched an action movie, while Group B watched an educational video related to the lecture topic. After 15 min, both groups attended a lecture on the gross anatomy of the heart, which was delivered by a qualified anatomist. At the end of the lecture, their understanding of the material was measured through a post-lecture test using ten vetted multiple choice true/false questions.

Results

Group B's test scores were found to be significantly higher than Group A's (p > 0.001, t-stats [df] = -4.21 [252]).

Conclusion

This study concluded that the pre-lecture activity had successfully provided the students with some prior knowledge of the subject before they attended the lecture sessions. This finding was aligned with cognitive load theory, which describes a reduction in learners' cognitive load when prior knowledge is stimulated.

Aromatase inhibitors: The unexpected breast cancer treatment

OBJECTIVE

Suboptimal adherence with endocrine treatment for breast cancer is influenced by a number of factors but remains poorly understood. We sought to describe the prior knowledge about and expectations of breast cancer treatments among older women retrospecting on their diagnosis and treatment.

METHODS

Thematic analysis was used to systematically analyze data obtained with face-to-face, open-ended interviews conducted with 54 women who had filled at least one prescription for an aromatase inhibitor. The average age was 71.9 (65-93) years at diagnosis.

RESULTS

Three salient themes were described: the sources of information on which preknowledge and expectations surrounding treatment were founded, and two phases of treatment, primary (surgery, chemotherapy and radiation therapy) and anti-hormonal. The main source of information was from family and friends who had been treated for cancer. These peers reported both positive and negative experiences and in many cases contributed to the women having some degree of misinformation. A foundational knowledge of primary treatments was evident (necessity, duration, intensity, side-effects) and that receiving one or more treatments was needed. Compared to primary treatments, anti-hormonal treatment (AHT) was unexpected, the women knew less about it, and felt comparatively under-prepared for this treatment.

CONCLUSIONS

The transition from primary treatments to adjuvant AHT therapy with receiving a prescription for an aromatase inhibitor caught many participants off guard. Our findings elucidate areas to enhance clinical practice, expand the research agenda to more thoroughly explore AHT information and design of an age-appropriate supportive intervention to improve continuation with AHT.

Knowledge-aware attention network for protein-protein interaction extraction

Protein-protein interaction (PPI) extraction from published scientific literature provides additional support for precision medicine efforts. However, many of the current PPI extraction methods need extensive feature engineering and cannot make full use of the prior knowledge in knowledge bases (KBs). KBs contain huge amounts of structured information about entities and relationships, therefore play a pivotal role in PPI extraction. This paper proposes a knowledge-aware attention network (KAN) to fuse prior knowledge about protein-protein pairs and context information for PPI extraction. The proposed model first adopts a diagonal-disabled multi-head attention mechanism to encode context sequence along with knowledge representations learned from KBs. Then a novel multi-dimensional attention mechanism is used to select the features that can best describe the encoded context. Experiment results on the BioCreative VI PPI dataset show that the proposed approach could acquire knowledge-aware dependencies between different words in a sequence and lead to a new state-of-the-art performance.

Do Alzheimer's Disease Patients Benefit From Prior-Knowledge in Associative Recognition Memory?

OBJECTIVES

Although the influence of prior knowledge on associative memory in healthy aging has received great attention, it has never been studied in Alzheimer's disease (AD). This study aimed at assessing whether AD patients could benefit from prior knowledge in associative memory and whether such benefit would be related to the integrity of their semantic memory.

METHODS

Twenty-one AD patients and 21 healthy older adults took part in an associative memory task using semantically related and unrelated word pairs and were also submitted to an evaluation of their semantic memory.

RESULTS

While participants of both groups benefited from semantic relatedness in associative discrimination, related pairs recognition was significantly predicted by semantic memory integrity in healthy older adults only.

CONCLUSIONS

We suggest that patients benefitted from semantic knowledge to improve their performance in the associative memory task, but that such performance is not related to semantic knowledge integrity evaluation measures because the two tasks differ in the way semantic information is accessed: in an automatic manner for the associative memory task, with automatic processes thought to be relatively preserved in AD, and in a controlled manner for the semantic knowledge evaluation, with controlled processes thought to be impaired in AD. (JINS, 2019, 25, 443-452).

Interplay between prior knowledge and communication mode on teaching effectiveness: Interpersonal neural synchronization as a neural marker

Teacher-student interaction allows students to combine prior knowledge with new information to develop new knowledge. It is widely understood that both communication mode and students' knowledge state contribute to the teaching effectiveness (i.e., higher students' scores), but the nature of the interplay of these factors and the underlying neural mechanism remain unknown. In the current study, we manipulated the communication modes (face-to-face [FTF] communication mode/computer-mediated communication [CMC] mode) and prior knowledge states (with vs. without) when teacher-student dyads participated in a teaching task. Using functional near-infrared spectroscopy, the brain activities of both the teacher and student in the dyads were recorded simultaneously. After teaching, perceived teacher-student interaction and teaching effectiveness were assessed. The behavioral results demonstrated that, during teaching with prior knowledge, FTF communication improved students' academic performance, as compared with CMC. Conversely, no such effect was found for teaching without prior knowledge. Accordingly, higher task-related interpersonal neural synchronization (INS) in the left prefrontal cortex (PFC) was found in the FTF teaching condition with prior knowledge. Such INS mediated the relationship between perceived interaction and students' test scores. Furthermore, the cumulative INS in the left PFC could predict the teaching effectiveness early in the teaching process (around 25-35 s into the teaching task) only in FTF teaching with prior knowledge. These findings provide insight into how the interplay between the communication mode and students' knowledge state affects teaching effectiveness. Moreover, our findings suggest that INS could be a possible neuromarker for dynamic evaluation of teacher-student interaction and teaching effectiveness.

Iterative spatial fuzzy clustering for 3D brain magnetic resonance image supervoxel segmentation

BACKGROUND

Although supervoxel segmentation methods have been employed for brain Magnetic Resonance Image (MRI) processing and analysis, due to the specific features of the brain, including complex-shaped internal structures and partial volume effect, their performance remains unsatisfactory.

NEW METHODS

To address these issues, this paper presents a novel iterative spatial fuzzy clustering (ISFC) algorithm to generate 3D supervoxels for brain MRI volume based on prior knowledge. This work makes use of the common topology among the human brains to obtain a set of seed templates from a population-based brain template MRI image. After selecting the number of supervoxels, the corresponding seed template is projected onto the considered individual brain for generating reliable seeds. Then, to deal with the influence of partial volume effect, an efficient iterative spatial fuzzy clustering algorithm is proposed to allocate voxels to the seeds and to generate the supervoxels for the overall brain MRI volume.

RESULTS

The performance of the proposed algorithm is evaluated on two widely used public brain MRI datasets and compared with three other up-to-date methods.

CONCLUSIONS

The proposed algorithm can be utilized for several brain MRI processing and analysis, including tissue segmentation, tumor detection and segmentation, functional parcellation and registration.

Semantic relation extraction aware of N-gram features from unstructured biomedical text

Semantic relation extraction is a crucial step of automatically constructing a knowledge graph from unstructured biomedical text. Many real-world applications can benefit from it. As unsupervised relation extraction approaches, generative probabilistic models, Rel-LDA and Type-LDA, are receiving more attention in recent years. However, these two models inherit the bag-of-word assumption of the standard LDA model, which disable the exploitation of more distinguishable n-gram features. To overcome this limitation, two alternative models, named as Rel-TNG and Type-TNG, are proposed with the help of Topic N-Grams (TNG) model in this study, and collapsed Gibbs sampling algorithm is utilized for inference. Extensive experimental results on GENIA and EPI corpora indicate that Rel-TNG and Type-TNG models have similar performance with their unigram counterparts, but Rel-TNG and Type-TNG models outperform Rel-LDA and Type-LDA models when prior knowledge is available.

Linguistic entrenchment: Prior knowledge impacts statistical learning performance

Statistical Learning (SL) is typically considered to be a domain-general mechanism by which cognitive systems discover the underlying statistical regularities in the input. Recent findings, however, show clear differences in processing regularities across modalities and stimuli as well as low correlations between performance on visual and auditory tasks. Why does a presumably domain-general mechanism show distinct patterns of modality and stimulus specificity? Here we claim that the key to this puzzle lies in the prior knowledge brought upon by learners to the learning task. Specifically, we argue that learners' already entrenched expectations about speech co-occurrences from their native language impacts what they learn from novel auditory verbal input. In contrast, learners are free of such entrenchment when processing sequences of visual material such as abstract shapes. We present evidence from three experiments supporting this hypothesis by showing that auditory-verbal tasks display distinct item-specific effects resulting in low correlations between test items. In contrast, non-verbal tasks - visual and auditory - show high correlations between items. Importantly, we also show that individual performance in visual and auditory SL tasks that do not implicate prior knowledge regarding co-occurrence of elements, is highly correlated. In a fourth experiment, we present further support for the entrenchment hypothesis by showing that the variance in performance between different stimuli in auditory-verbal statistical learning tasks can be traced back to their resemblance to participants' native language. We discuss the methodological and theoretical implications of these findings, focusing on models of domain generality/specificity of SL.

Prior knowledge modulates the neural substrates of encoding and retrieving naturalistic events at short and long delays

Congruence with prior knowledge and incongruence/novelty have long been identified as two prominent factors that, despite their opposing characteristics, can both enhance episodic memory. Using narrative film clip stimuli, this study investigated these effects in naturalistic event memories - examining behaviour and neural activation to help explain this paradox. Furthermore, we examined encoding, immediate retrieval, and one-week delayed retrieval to determine how these effects evolve over time. Behaviourally, both congruence with prior knowledge and incongruence/novelty enhanced memory for events, though incongruent events were recalled with more errors over time. During encoding, greater congruence with prior knowledge was correlated with medial prefrontal cortex (mPFC) and parietal activation, suggesting that these areas may play a key role in linking current episodic processing with prior knowledge. Encoding of increasingly incongruent events, on the other hand, was correlated with increasing activation in, and functional connectivity between, the medial temporal lobe (MTL) and posterior sensory cortices. During immediate and delayed retrieval the mPFC and MTL each demonstrated functional connectivity that varied based on the congruence of events with prior knowledge; with connectivity between the MTL and occipital regions found for incongruent events, while congruent events were associated with functional connectivity between the mPFC and the inferior parietal lobules and middle frontal gyri. These results demonstrate patterns of neural activity and connectivity that shift based on the nature of the event being experienced or remembered, and that evolve over time. Furthermore, they suggest potential mechanisms by which both congruence with prior knowledge and incongruence/novelty may enhance memory, through mPFC and MTL functional connectivity, respectively.

Language Skills in Classical Chinese Text Comprehension

This study used both quantitative and qualitative methods to explore the role of lower- and higher-level language skills in classical Chinese (CC) text comprehension. A CC word and sentence translation test, text comprehension test, and questionnaire were administered to 393 Secondary Four students; and 12 of these were randomly selected to participate in retrospective interviews. The findings revealed that students' CC reading performance was unsatisfactory with respect to both lower- and text-level comprehension. Among the different factors examined, the most crucial to CC reading comprehension was lower-level reading skill. Owing to students' weak lower-level reading skills, participants relied heavily on contextual clues when reading in CC. The implications of these findings for understanding factors that contribute to CC reading comprehension, and for planning effective instruction to enhance students' CC reading competence are discussed.