No abstract

This is a repository copy of Predicting peak load of bus routes with supply optimization and scaled Shepard interpolation: A newsvendor model.

e14534 Background: Nivolumab is a PD-1 inhibitor that is FDA approved for treatment of chemotherapy refractory advanced NSCLC. The current standard clinical approach to evaluating tumor response is sub-optimal in evaluating clinical benefit from immunotherapy drugs. Our study aims to explore whether changes in radiomic features of the tumor between baseline and 2-week post-treatment CT scans can predict treatment response. Methods: 41 NSCLC patients treated with nivolumab were included in this study. 22 patients with pre- and post-nivolumab CT scans were used as a learning set and the remaining 19 for independent testing. Patients who did not receive nivolumab after 2 cycles due to lack of response or progression as per RECIST were classified as ‘non-responders’, and patients who had radiological response as per RECIST, or stable disease as per RECIST and clinical improvement were classified as ‘responders’. Lung nodules on pre-treatment CT scans were annotated with 3D SLICER software by a radiologist. 312 texture features of lung nodules were extracted and investigated in the study. The percent difference of each extracted feature was calculated based on the baseline and 2 week post-therapy CT scan. In the learning set, the six features that most significantly changed between baseline and post-treatment scans and also maximally differentially expressed between responders and non-responders were identified. Unsupervised clustering was applied on the set of 6 features for the 19 patients in the test set to predict which patients did and did not respond. Results: The top 6 features predictive of response corresponded to the Haralick, Gabor and Laws texture families. Unsupervised clustering yielded an accuracy of 78.95%. Conclusions: Our results suggest that changes in certain radiomic texture features between baseline and post-treatment CT scans following nivolumab could identify early clinical response to treatment. Additional validation of these novel quantitative imaging based approaches is warranted to accurately define clinical benefit from immunotherapy.

Understanding the neural mechanisms of working memory has been a long-standing Neuroscience goal. Bump attractor models have been used to simulate persistent activity generated in the prefrontal cortex during working memory tasks and to study the relationship between activity and behavior. How realistic the assumptions of these models are has been a matter of debate. Here, we relied on an alternative strategy to gain insights into the computational principles behind the generation of persistent activity and on whether current models capture some universal computational principles. We trained Recurrent Neural Networks (RNNs) to perform spatial working memory tasks and examined what aspects of RNN activity accounted for working memory performance. Furthermore, we compared activity in fully trained networks and immature networks, achieving only imperfect performance. We thus examined the relationship between the trial-to-trial variability of responses simulated by the network and different aspects of unit activity as a way of identifying the critical parameters of memory maintenance. Properties that spontaneously emerged in the artificial network strongly resembled persistent activity of prefrontal neurons. Most importantly, these included drift of network activity during the course of a trial that was causal to the behavior of the network. As a consequence, delay period firing rate and behavior were positively correlated, in strong analogy to experimental results from the prefrontal cortex. These findings reveal that delay period activity is computationally efficient in maintaining working memory, as evidenced by unbiased optimization of parameters in artificial neural networks, oblivious to the properties of prefrontal neurons.

To study the influence mechanism of dedicated bus lanes on the urban road network, this paper proposes a novel analytical model of macroscopic fundamental diagram (MFD) and passenger macroscopic fundamental diagram (p-MFD) and the corresponding indicators based on MFD and p-MFD to evaluate the operation of the network. Taking the grid network as an example, this paper collects traffic flow to calibrate the developed MFD and p-MFD and evaluates the network performance under different proportions of dedicated bus lanes. The simulation results show that the larger the proportion of dedicated bus lanes, the greater the impact on the rising section and the stable section of MFD and the descending section and post-stable section of p-MFD. Further analysis for the sensitivity of simulation experiments found that the strategy of setting dedicated bus lanes will improve the efficiency of vehicle and passenger transport when the road network is in a smooth state and ensure the continuous output of passengers when the network is in a congested state.