Code Locations Research Articles

Auditory competition and stimulus selection across spatial locations from midbrain to forebrain in barn owls

Barn owls enable investigation of neural mechanisms underlying stimulus selection of concurrent stimuli. The audio-visual space map in the optic tectum (OT), avian homologue of the superior colliculus, encodes relative strength of concurrent auditory stimuli through spike response rate and interneuronal spike train synchrony (STS). Open questions remain regarding stimulus selection in downstream forebrain regions lacking topographic coding of auditory space, including the functional consequences of interneuronal STS on interregional signaling. To this end, we presented concurrent stimuli at different locations and manipulated relative strength while simultaneously recording neural responses from OT and its downstream thalamic target, nucleus rotundus (nRt), in awake barn owls of both sexes. Results demonstrated that broadly spatially tuned nRt units exhibit different spike response patterns to competition depending on spatial tuning preferences. Modeling suggests nRt units integrate convergent inputs from distant locations across midbrain map regions. Additionally, STS within nRt reflects the temporal properties of the strongest stimulus. Furthermore, interregional STS between OT and nRt was strongest when spatial tuning overlap between units across regions was large and when the strongest stimulus location during competition was favorable for units in both regions. Additionally, though gamma oscillations synthesized within OT are weakly propagated within nRt, average gamma power across regions correlates with strength of interregional STS. Overall, we demonstrate that nRt integrates inputs across distant areas of OT, retains spatial information through differences in strength of inputs from various locations of the midbrain map across neurons, and prioritizes coding of identity features to the strongest sound.Significance StatementThe brain strategically selects and preferentially processes salient stimuli. A critical function to this process involves transferring salient information across regions that may exhibit drastic transformations in coding schemes. Our study in barn owls investigates bottom-up signaling between the midbrain space map and its downstream thalamic target, which lacks spatial topography as also observed in mammalian auditory forebrain regions to elucidate general mechanisms underlying how spatial location information and other properties of the strongest sound are relayed between regions. Results show that the thalamus integrates neural responses widely across the midbrain map, retains coding of spatial location through varying strength of inputs of the map across neurons, and prioritizes further coding of identity features only to the strongest sound.

Distinct Neuron Types Contribute to Hybrid Auditory Spatial Coding.

Neural decoding is a tool for understanding how activities from a population of neurons inside the brain relate to the outside world and for engineering applications such as brain-machine interfaces. However, neural decoding studies mainly focused on different decoding algorithms rather than different neuron types which could use different coding strategies. In this study, we used two-photon calcium imaging to assess three auditory spatial decoders (space map, opponent channel, and population pattern) in excitatory and inhibitory neurons in the dorsal inferior colliculus of male and female mice. Our findings revealed a clustering of excitatory neurons that prefer similar interaural level difference (ILD), the primary spatial cues in mice, while inhibitory neurons showed random local ILD organization. We found that inhibitory neurons displayed lower decoding variability under the opponent channel decoder, while excitatory neurons achieved higher decoding accuracy under the space map and population pattern decoders. Further analysis revealed that the inhibitory neurons' preference for ILD off the midline and the excitatory neurons' heterogeneous ILD tuning account for their decoding differences. Additionally, we discovered a sharper ILD tuning in the inhibitory neurons. Our computational model, linking this to increased presynaptic inhibitory inputs, was corroborated using monaural and binaural stimuli. Overall, this study provides experimental and computational insight into how excitatory and inhibitory neurons uniquely contribute to the coding of sound locations.

Lateral Entorhinal Cortex Dysfunction in Alzheimer's Disease Mice.

In Alzheimer's disease (AD), the formation of amyloid beta and neurofibrillary tangles (NFTs) leads to neuronal loss in entorhinal cortex (EC), a crucial brain region involved in memory and navigation. These pathological changes are concurrent with the onset of memory-related issues in AD patients with symptoms of forgetfulness such as misplacing items, disorientation in familiar environments etc. The lateral EC (LEC) is associated with non-spatial memory processing including object recognition. Since in LEC, neurons fire in response to objects (object cells) and at locations previously occupied by objects (trace cells), pathology in this region could lead to dysfunction in object location coding. In this paper we show that a transgenic mouse model, EC-App/Tau, which expresses both APP and tau primarily in the EC region, have deficits in LEC-specific memory tasks. Using in vivo single-unit electrophysiology recordings we show that the LEC neurons are hyperactive with low information content and high sparsity compared to the controls indicating poor firing fidelity. We finally show that object cells and trace cells fire less precisely in the EC-App/Tau mice compared to controls indicating poor encoding of objects. Overall, we show that AD pathology causes erratic firing of LEC neurons and object coding defects leading to LEC-specific memory impairment.

SeSMR: Secure and Efficient Session-based Multimedia Recommendation in Edge Computing

Session-based multimedia recommendation in edge computing remains an important issue for boosting the utilization of services since service composition has increasingly attracted attention. Existing session-based recommendations (SBRs) model the session sequence with multilevel feature extraction in graph neural networks (GNNs). However, multilevel feature extraction in disentangled graph neural networks causes over-smoothing and privacy leakage. To address the aforementioned problems, Secure and Efficient Session-based Multimedia Recommendation (SeSMR) model is proposed. In the proposed SeSMR model, based on BGV homomorphic encryption, a ciphertext training submodel is proposed to address the privacy leakage, ensuring the security in session-based recommendation. Furthermore, based on the reinforcement of feature activation, a residual attention mechanism is proposed to mitigate over-smoothing while maintaining the independence of multiple features. Finally, based on location coding, a soft attention mechanism is proposed to improve the recommendation accuracy, by introducing the position difference information between items into intra-session and inter-session scenarios. Experiments demonstrate that both Recall and MRR metrics exhibit nearly 2%∼5% improvement.

Scene semantics affects allocentric spatial coding for action in naturalistic (virtual) environments

Interacting with objects in our environment requires determining their locations, often with respect to surrounding objects (i.e., allocentrically). According to the scene grammar framework, these usually small, local objects are movable within a scene and represent the lowest level of a scene’s hierarchy. How do higher hierarchical levels of scene grammar influence allocentric coding for memory-guided actions? Here, we focused on the effect of large, immovable objects (anchors) on the encoding of local object positions. In a virtual reality study, participants (n = 30) viewed one of four possible scenes (two kitchens or two bathrooms), with two anchors connected by a shelf, onto which were presented three local objects (congruent with one anchor) (Encoding). The scene was re-presented (Test) with 1) local objects missing and 2) one of the anchors shifted (Shift) or not (No shift). Participants, then, saw a floating local object (target), which they grabbed and placed back on the shelf in its remembered position (Response). Eye-tracking data revealed that both local objects and anchors were fixated, with preference for local objects. Additionally, anchors guided allocentric coding of local objects, despite being task-irrelevant. Overall, anchors implicitly influence spatial coding of local object locations for memory-guided actions within naturalistic (virtual) environments.

Gender based disparities in Medicare physician reimbursement persist across years and specialty.

The gender pay gap is wide in medicine but the extent of this disparity across specialties and over time have not been elucidated. Here we evaluate differences in Medicare reimbursement between men and women physicians over time and by specialty, controlling for physician and practice characteristics. The Centers for Medicare & Medicaid Services Payment Data was used to determine total reimbursements and number of services submitted by physicians practicing in the US between 2013 and 2019. Data from the American Community Survey (ACS) were used to determine average income, unemployment rates, poverty rates, income, and educational attainment levels by zip code for each physician's practice location. Among the 3,831,504 physicians included in this analysis from 2013-2019, 2,712,545 (70.8%) were men and 1,118,859 (29.2%) were women. Overall, men received more in Medicare reimbursements ($58,815 ± $104,772 vs. $32,205 ± $60,556, p<0.001) and billed more services (864 ± 1,780 vs. 505 ± 1,007, p<0.001) compared to women. The median Medicare reimbursement for men decreased from 2013 to 2019 from $59,710 to $57,874, while the median Medicare reimbursement for women increased from $30,575 to $33,456. Men were reimbursed more than women across all specialties with the greatest disparity in procedure-heavy specialties. The specialties with the highest difference in median Medicare reimbursement between men and women were ophthalmology ($99,452), dermatology ($84,844), cardiology ($64,112), nephrology ($62,352), and pulmonary medicine ($47,399). In linear regression models controlling for calendar year, years of experience, total number of services, and ACS zip-code-level variables, men received a higher amount of Medicare reimbursement in all specialties, as compared to women (p<0.01 for all). The percentage of top earning men (range: 65.0%-99.5%) surpassed the proportion of men in each specialty (range: 46.1%-94.6%), except public health and preventive medicine. Women physicians continue to receive lower total Medicare reimbursements than men physicians, particularly in procedure-heavy specialties. Lower clinical volume and fewer procedural services among women physicians partially contribute to the disparities in reimbursement.

Multi‐objective optimization‐based and fault localization‐oriented test case generation for novice programs

SummaryOnline judgment (OJ) systems are capable of evaluating program results by automatically executing test cases, significantly improving the efficiency of traditional guidance approaches. Moreover, existing studies attempt to assist novices through automated fault localization techniques to provide feedback to novices, which can help them quickly find the location of faulty statements. Among them, spectrum‐based fault localization (SBFL) techniques have been widely used for their lightweight and efficiency, which only requires coverage information and test results of test cases to conduct fault localization. However, manually constructing high‐quality test cases for a large number of OJ questions is tough work to complete. To solve this problem, we propose the novice program‐oriented Multi‐Objective Optimization‐Based Fault Localization‐Oriented Test Case Generation (MFTCG) for automatically generating test inputs. Specifically, we use multi‐objective optimization algorithms to evolve the test case in terms of both fault localization and faulty code detection capability. We conduct experiments with 8911 programs from the well‐known public OJ platform AtCoder. The results show that our proposed approach MFTCG can achieve the best fault localization performance compared with existing automated test case generation approaches in most cases and can achieve the similar faulty code detection capability compared to manually designed test cases.

Mapping the tree damage index: A case study on the green lane in the urban area of Banda Aceh City

Climate change mitigation efforts implemented by the Government of Banda Aceh by building green lanes with trees. The level of the tree’s ability to absorb carbon is highly correlated with the vitality of the tree itself. The health status of a tree can be known by the tree’s damage level. The aim of the research is to provide information on how much damage to trees and the degree of severity. The research was carried out on green lanes on urban streets in Banda Aceh City as many as 8 lines. The object of research is vegetation with the tree growth phase category, and tree damage index assessment using the forest health monitoring method with indicators of tree damage conditions with parameters of damage location code, damage type code, and damage level code. Based on the results, most of the trees suffered light damage. Of the total number of trees observed, there were 1797 individuals, 252 trees were in good health, 1494 were lightly damaged, 40 trees were medium damaged and 11 trees were heavily damaged. The type of damage experienced was dominated by 46% open wound, 32% cancer, 11% brum (excessive branching) and 11% the rest with other kinds of damage.

D2D Self Organization in IOT via Triple Modular Redundancy Based MDS Code

Device-to-Device (D2D) communication is a key enabling technology to replace base station-to-device(B2D) communication. D2D with IoT realization has spectral and energy-efficient features for on-time quality data transfer. However, the rapid growth of such networks and the dynamic quality of service demands of user equipment require robust and efficient service quality in disasters, outages, and adversarial conditions. To address this challenge, this paper proposes a triple modular redundancy based on the maximum distance separable (TMR-MDS) technique has been proposed, which modifies the repair bandwidth by utilizing more storage occupation and improving the D2D link availability in the system. The proposed TMR-MDS technique will be used, which is composed of two phases: trust management modules and triple modular redundancy storage modules. As part of the initial stage of trust management, a fuzzy logic model is employed to determine the user’s trust value after the user’s behaviour is examined. Second, using the majority voting scheme to improve triple modular redundancy storage. Furthermore, if any problem occurs in the node, then it will self-organize. MATLAB is used to evaluate the performance of the proposed system. The overall system performance is evaluated in terms of the network’s trust level, and the communication cost of the proposed strategy. As per the experimental results, the proposed TMR-MDS method has a better performance of 12.67%, 23.68%, and 25.87% than hierarchical bi-partite, double replication MDS, and unequal repairing locality code.

Image inpainting method based on AU-GAN

Image inpainting refers to the process of filling in missing regions or removing objects, and has broad application prospects. The rapid development of deep learning has led to new technological breakthroughs in image repair technology, continuously improving the quality of image inpainting. However, when we inpaint large missing regions, the texture and structural features of the image cannot be comprehensively utilized. This leads to blurry images. To solve this problem, we propose an improved dual-stream U-Net algorithm that adds an attention mechanism to the two U-Net networks known as a dual AU-Net network to improve the texture details of the image. In addition, the location code (LC) of damaged regions is added to the network to guide network repair and accelerate the network convergence speed. Least squares GAN (LSGAN) loss is added to the generator’s adversarial network to capture more content details and enhance training stability. The PSNR is 33.93 and the SSIM is 0.931 in the CelebA and Paris datasets. This method has been proven effective when compared to other methods.

Vertical stratification and seasonality of fecal indicator bacteria in New York City playground sandboxes

Sandboxes in public play spaces afford a crucial opportunity for urban children to engage in naturalistic play that fosters development of cognitive, social, and motor skills. As open pits, sandboxes in New York City public playgrounds are potentially exposed to fecal inputs from various sources, including wild and domestic animals. A longitudinal study of thirteen sandboxes located in public playgrounds on the east side of Manhattan reveals ubiquity of the fecal indicator bacteria enterococci and Escherichia coli through all seasons. The highest concentrations of bacteria occur in surface sand (n = 42; mean enterococci 230 MPN/g and E. coli 182 MPN/g dry weight), with significantly lower levels at depths below the surface (n = 35; mean enterococci 21 MPN/g and E. coli 12 MPN/g dry weight), a stratification consistent with fecal loading at the surface. Generalized linear mixed models indicate that sand depth (surface vs. underlayers) is the most influential variable affecting bacterial levels (P <0.001 for both enterococci and E. coli), followed by sampling season (P <0.001 for both). Bacterial concentrations do not vary significantly as a function of playground location or ZIP code within the study area. Children’s exposure while playing in sandboxes likely reaches 105 enterococci and 104E. coli in a typical play period. Microbial source tracking to identify fecal hosts reveals dog, bird, and human biomarkers in low concentrations. Open sandbox microcosms installed at ground level in the urban environment of Manhattan are fouled by enterococci and E. coli within two weeks, while adjacent closed microcosms exhibit no fecal contamination over a 33-day sampling period. Collectively, our results indicate that increasing the frequency of sand refills and covering sandboxes during times of disuse would be straightforward management strategies to mitigate fecal contamination in playground sandboxes.

A novel nomogram and recursive partitioning analysis for predicting cancer-specific survival of patients with subcutaneous leiomyosarcoma

Accurately predicting prognosis subcutaneous leiomyosarcoma (LMS) is crucial for guiding treatment decisions in patients. The objective of this study was to develop prediction models for cancer-specific survival (CSS) in patients with subcutaneous LMS. The collected cases of diagnosed subcutaneous LMS were randomly divided into a training cohort and a validation cohort at a 6:4 ratio based on tumor location and histological code. The X-tile program was utilized to determine the optimal cutoff points for age index. Univariate and Cox multivariate regression analyses were conducted to identify independent risk factors for subcutaneous LMS patients. Nomograms were constructed to predict CSS, and their performance was assessed using C-index and calibration plots. Additionally, a decision tree model was established using recursive partitioning analysis to determine the total score for CSS prediction in subcutaneous LMS patients based on the nomogram model. A total of 1793 patients with subcutaneous LMS were found. X-tile software divides all patients into ≤ 61 years old, 61–82 years old, and ≥ 82 years old. The most important anatomical sites were the limbs (including the upper and lower limbs, 48.0%). Only 6.2% of patients received chemotherapy, while 44% had a history of radiotherapy and 92.9% underwent surgery. The independent risk factors for patients with subcutaneous LMS were age, summary stage, grade, and surgery. CSS was significantly decreased in patients with distant metastases, which showed the highest independent risk predictor (HR 4.325, 95% CI 3.010–6.214, p < 0.001). The nomogram prediction model of LMS was constructed based on four risk factors. The C-index for this model was 0.802 [95% CI 0.781–0.823] and 0.798 [95% CI 0.768–0.829]. The training cohort's 3-, 5-, and 10-year AUCs for CSS in patients with subcutaneous LMS were 0.833, 0.830, and 0.859, and the validation cohort's AUC for predicting CSS rate were 0.849, 0.830, and 0.803, respectively. Recursive segmentation analysis divided patients into 4 risk subgroups according to the total score in the nomogram, including low-risk group < 145, intermediate-low-risk group ≥ 145 < 176, intermediate-high-risk group ≥ 176 < 196, and high-risk group ≥ 196; The probability of the four risk subgroups is 9.1%, 34%, 49%, and 79% respectively. In this retrospective study, a novel nomogram or corresponding risk classification system for patients with subcutaneous LMS were developed, which may assist clinicians in identifying high-risk patients and in guiding the clinical decision.

TSAGNN: Temporal link predict method based on two stream adaptive graph neural network

Temporal link prediction based on graph neural networks has become a hot spot in the field of complex networks. To solve the problems of the existing temporal link prediction methods based on graph neural networks do not consider the future time-domain features and spatial-domain features are limited used, this paper proposes a novel temporal link prediction method based on two streams adaptive graph neural networks. Firstly, the network topology features are extracted from the micro, meso, and middle perspectives. Combined with the adaptive mechanism of convolution and self-attention, the preprocessing of the feature extraction is more effective; Secondly, an extended bi-directional long short-term memory network is proposed, which uses graph convolution to process topological features, and recursively learns the state vectors of the target snapshot by using the future time-domain information and the past historical information; Thirdly, the location coding is replaced by the time-coding for the transformer mechanism, so that past information and future information can be learned from each other, and the time-domain information of the network can be further mined; Finally, a novel two-stream network framework is proposed, which combines the processing results of point features and edge features. The experimental results on 9 data sets show that the proposed method has a better prediction effect and better robustness than the classical graph neural network methods.

Using NEMSIS to report annual air medical transport statistics

The objective of this study was to use the National Emergency Medical Services Information System (NEMSIS) dataset to generate national air medical transport statistics. Retrospective review of the 2021 NEMSIS dataset to identify all air medical transfers, both fixed- and rotor-wing. Transfers where then subcategorized into interfacility and scene responses. Frequencies for each category were generated and reported. A total of 317,267 air medical transfers were completed in 2021. These included 19,421 (6 %) with missing incident location code data. Of the 297,706 transfers with valid location codes, 208,689 (70%) were interfacility transfers, and 89,016 (30%) were scene responses. These preliminary results are consistent with other national estimates and achieve representation of all states and territories. Future work will include longitudinal analysis of NEMSIS datasets and direct survey of transport programs to establish long-term reliability.

Sensitive Data Privacy Protection of Carrier in Intelligent Logistics System

An intelligent logistics system is a production system based on the Internet of Things (IoT), and the logistics information of humans has a high degree of privacy. However, the current intelligent logistics system only protects the privacy of shippers and consignees, without any privacy protection for carriers, which will not only cause carriers’ privacy leakage but also indirectly or directly affect the logistics efficiency. It is particularly worth noting that solving this problem requires one to consider the balance between privacy protection and operational visibility. So, the local privacy protection algorithm ϵ-L_LDP for carriers’ multidimensional numerical sensitive data and ϵ-LT_LDP for carrier location sensitive data are proposed. For ϵ-L_LDP, firstly, a personalized and locally differentiated privacy budgeting approach is used. Then, the multidimensional data personalization perturbation mechanism algorithm L-PM is designed. Finally, the multidimensional data are perturbed using L-PM. For ϵ-LT_LDP, firstly, the location area is matrix-partitioned and quadtree indexed, and the location data are indexed according to the quadtree to obtain the geographic location code in which it is located. Secondly, the personalized random response perturbation algorithm L-RR for location trajectory data is also designed. Finally, the L-RR algorithm is used to implement the perturbation of geolocation-encoded data. Experiments are conducted using real and simulated datasets, the results show that the ϵ-L_LDP algorithm and ϵ-LT_LDP algorithm can better protect the privacy information of carriers and ensure the availability of carrier data during the logistics process. This effectively meets the balance between the privacy protection and operational visibility of the intelligent logistics system.

Eye gaze is not unique: The reversed congruency effect on gaze and tongue targets.

In the spatial Stroop task, an arrow target produces a spatial Stroop effect, whereas a gaze target elicits a reversed congruency effect. The reversed congruency effect has been explained by the unique attentional mechanisms of eye gaze. However, recent studies have shown that not only gaze but arrow targets produced a reversed congruency effect when embedded in a complex background. The present study investigated whether non-gaze targets produce a reversed congruency effect. In Experiments 1 and 2, we used the tongue, which is not commonly used to indicate spatial directions in daily life, as a target in the spatial Stroop task, in addition to the conventional gaze and arrows. In Experiment 3, we used arrow stimuli embedded in a complex background as a target. Participants judged the left/right direction of the target presented in the left or right visual field. Although arrow and gaze targets replicated previous findings (spatial Stroop and reversed congruency effect, respectively), the tongue target produced a reversed congruency effect (Experiments 1 and 2). The spatial Stroop effect of arrow targets disappeared when they were in a complex background (Experiment 3). These results are inconsistent with previous accounts emphasising the unique status of eye gaze. We propose that temporal decay of the location code and response inhibition are responsible for the reversal of spatial interference.

HAM-Transformer: A Hybrid Adaptive Multi-Scaled Transformer Net for Remote Sensing in Complex Scenes

The quality of remote sensing images has been greatly improved by the rapid improvement of unmanned aerial vehicles (UAVs), which has made it possible to detect small objects in the most complex scenes. Recently, learning-based object detection has been introduced and has gained popularity in remote sensing image processing. To improve the detection accuracy of small, weak objects in complex scenes, this work proposes a novel hybrid backbone composed of a convolutional neural network and an adaptive multi-scaled transformer, referred to as HAM-Transformer Net. HAM-Transformer Net firstly extracts the details of feature maps using convolutional local feature extraction blocks. Secondly, hierarchical information is extracted, using multi-scale location coding. Finally, an adaptive multi-scale transformer block is used to extract further features in different receptive fields and to fuse them adaptively. We implemented comparison experiments on a self-constructed dataset. The experiments proved that the method is a significant improvement over the state-of-the-art object detection algorithms. We also conducted a large number of comparative experiments in this work to demonstrate the effectiveness of this method.

A fine-grained causality extraction model incorporating relative location coding

A fine-grained causality extraction model incorporating relative location coding

SLBRIN: A Spatial Learned Index Based on BRIN

The spatial learned index constructs a spatial index by learning the spatial distribution, which performs a lower cost of storage and query than the spatial indices. The current update strategies of spatial learned indices can only solve limited updates at the cost of query performance. We propose a novel spatial learned index structure based on a Block Range Index (SLBRIN for short). Its core idea is to cooperate history range and current range to satisfy a fast spatial query and efficient index update simultaneously. SLBRIN deconstructs the update transaction into three parallel operations and optimizes them based on the temporal proximity of spatial distribution. SLBRIN also provides the spatial query strategy with the spatial learned index and spatial location code, including point query, range query and kNN query. Experiments on synthetic and real datasets demonstrate that SLBRIN clearly outperforms traditional spatial indices and state-of-the-art spatial learned indices in the cost of storage and query. Moreover, in the simulated real-time update scenario, SLBRIN has the faster and more stable query performance while satisfying efficient updates.

Mnemosyne: Privacy-Preserving Ride Matching With Collusion-Resistant Driver Exclusion

Ride-Hailing Service (RHS) has drawn plenty of attention as it provides transportation convenience for riders and financial incentives for drivers. Despite these benefits, riders risk the exposure of sensitive location data during ride requesting to an untrusted Ride-Hailing Service Provider (RHSP). Our motivation arises from <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">repetitive matching, i.e., the same driver is repetitively assigned to the same rider</i> . Meanwhile, we introduce a driver exclusion function to protect riders' location privacy. Existing work on privacy-preserving RHS overlooks this function. While Secure <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$k$</tex-math></inline-formula> Nearest Neighbor (S <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$k$</tex-math></inline-formula> NN) facilitates efficient matching, the state-of-the-art neglects a collusion attack. To solve this problem, we formally define repetitive matching and strong location privacy, and propose Mnemosyne: privacy-preserving ride matching with collusion-resistant driver exclusion. We extend the simple integration of equality checking and item exclusion to a dynamic integration. We concatenate each prefix of an acceptable identity range to each location code when generating a ride request, i.e., secure mix index. We process each prefix of the driver identity to generate a ride response, i.e., a mix token. We build an indistinguishable Bloom-filter as an index to query the token. When matching riders with drivers, the colluding parties cannot distinguish identity prefixes from location codes. We build a prototype of Mnemosyne based on servers, smartphones, and a real-world dataset. Experimental results demonstrate that Mnemosyne outperforms existing work regarding strong location privacy and computational costs.