Subject-specific Template Research Articles

Shape-aware Segmentation of the Placenta in BOLD Fetal MRI Time Series.

Blood oxygen level dependent (BOLD) MRI time series with maternal hyperoxia can assess placental oxygenation and function. Measuring precise BOLD changes in the placenta requires accurate temporal placental segmentation and is confounded by fetal and maternal motion, contractions, and hyperoxia-induced intensity changes. Current BOLD placenta segmentation methods warp a manually annotated subject-specific template to the entire time series. However, as the placenta is a thin, elongated, and highly non-rigid organ subject to large deformations and obfuscated edges, existing work cannot accurately segment the placental shape, especially near boundaries. In this work, we propose a machine learning segmentation framework for placental BOLD MRI and apply it to segmenting each volume in a time series. We use a placental-boundary weighted loss formulation and perform a comprehensive evaluation across several popular segmentation objectives. Our model is trained and tested on a cohort of 91 subjects containing healthy fetuses, fetuses with fetal growth restriction, and mothers with high BMI. Biomedically, our model performs reliably in segmenting volumes in both normoxic and hyperoxic points in the BOLD time series. We further find that boundary-weighting increases placental segmentation performance by 8.3% and 6.0% Dice coefficient for the cross-entropy and signed distance transform objectives, respectively.

Fixed template network and dynamic template network: novel network designs for decoding steady-state visual evoked potentials

Objective. Decomposition methods are efficient to decode steady-state visual evoked potentials (SSVEPs). In recent years, the brain–computer interface community has also been developing deep learning networks for decoding SSVEPs. However, there is no clear evidence that current deep learning models outperform decomposition methods on the SSVEP decoding tasks. Many studies lacked the comparison with state-of-the-art decomposition methods in a fair environment. Approach. This study proposed a novel network design motivated by the works of decomposition methods. Fixed template network (FTN) and dynamic template network (DTN) are two novel networks combining the advantages of fixed templates and subject-specific templates. This study also proposed a data augmentation method for SSVEPs. This study compared the intra-subject classification performance of DTN and FTN with that of state-of-the-art decomposition methods on three public SSVEP datasets. Main results. The results show that both FTN and DTN achieved the suboptimal classification performance compared with state-of-the-art decomposition methods. Significance. Both network designs could enhance the decoding performance of SSVEPs, making them promising networks for improving the practicality of SSVEP-based applications.

CVEP Training Data Validation-Towards Optimal Training Set Composition from Multi-Day Data.

This paper investigates the effects of the repetitive block-wise training process on the classification accuracy for a code-modulated visual evoked potentials (cVEP)-based brain–computer interface (BCI). The cVEP-based BCIs are popular thanks to their autocorrelation feature. The cVEP-based stimuli are generated by a specific code pattern, usually the m-sequence, which is phase-shifted between the individual targets. Typically, the cVEP classification requires a subject-specific template (individually created from the user’s own pre-recorded EEG responses to the same stimulus target), which is compared to the incoming electroencephalography (EEG) data, using the correlation algorithms. The amount of the collected user training data determines the accuracy of the system. In this offline study, previously recorded EEG data collected during an online experiment with 10 participants from multiple sessions were used. A template matching target identification, with similar models as the task-related component analysis (TRCA), was used for target classification. The spatial filter was generated by the canonical correlation analysis (CCA). When comparing the training models from one session with the same session’s data (intra-session) and the model from one session with the data from the other session (inter-session), the accuracies were (94.84%, 94.53%) and (76.67%, 77.34%) for intra-sessions and inter-sessions, respectively. In order to investigate the most reliable configuration for accurate classification, the training data blocks from different sessions (days) were compared interchangeably. In the best training set composition, the participants achieved an average accuracy of 82.66% for models based only on two training blocks from two different sessions. Similarly, at least five blocks were necessary for the average accuracy to exceed 90%. The presented method can further improve cVEP-based BCI performance by reusing previously recorded training data.

Language as context: A case of early literacy practices in New Zealand and Sweden

ABSTRACT Globally, there has been increasing adaptation of curricula frameworks in early childhood education, providing overarching principles of practice rather than subject specific templates for teaching and learning. While such a movement is to be commended as supporting a socio-cultural approach in meeting each child’s unique social and cultural needs, the implementation of frameworks is not straightforward (Bateman 2022). By applying an ethnomethodological approach to child–teacher interactions, this article explores how early childhood curricula frameworks in Sweden and New Zealand are implemented in everyday talk-in-interaction between children and teachers. We use an ethnomethodological (EM) approach (Garfinkel 2002) and conversation analysis (CA) (Sacks, Schegloff, and Jefferson 1974) approach (EMCA) to situate language as context. This approach offers a move away from the broader perspective of context being a static environmental space, to context as co-constructed by the participants through their immediate interactions (Goodwin and Duranti 1992). Contextual resources involve the concrete social situations, background cultural knowledge, language, activity and situation types, participants’ knowledge about topics talked about and about each other and their interactional biographies (Linell 2009, 17).

Learning deep features for task-independent EEG-based biometric verification

Abstract Considerable interest has been recently devoted to the exploitation of brain activity as biometric identifier in automatic recognition systems, with a major focus on data acquired through electroencephalography (EEG). Several researches have in fact confirmed the presence of discriminative characteristics within brain signals recorded while performing specific mental tasks. Yet, to make EEG-based recognition appealing for practical applications, it would be highly advisable to investigate the existence and permanence of such distinctive traits while performing different mental tasks. In this regard, the present study evaluates the feasibility of performing task-independent EEG-based biometric recognition. A deep learning approach using siamese convolutional neural networks is employed to extract, from the considered EEG recordings, subject-specific template representations. An extensive set of experimental tests, performed on a multi-session database comprising EEG data acquired from 45 subjects while performing six different tasks, is employed to evaluate whether it is actually possible to verify the identity of a subject using brain signals, regardless the performed mental task.

Toward Optimizing VEMP: Calculating VEMP Inhibition Depth With a Generic Template.

Cervical vestibular evoked myogenic potentials (cVEMP) indirectly reveal the response of the saccule to acoustic stimuli through the inhibition of sternocleidomastoid muscle electromyographic response. VEMP inhibition depth (VEMPid) is a recently developed metric that estimates the percentage of saccular inhibition. VEMPid provides both normalization and better accuracy at low response levels than amplitude-normalized cVEMPs. Hopefully, VEMPid will aid in the clinical assessment of patients with vestibulopatholgy. To calculate VEMPid a template is needed. In the original method, a subject's own cVEMP was used as the template, but this method can be problematic in patients who do not have robust cVEMP responses. We hypothesize that a "generic" template, created by assembling cVEMPs from healthy subjects, can be used to compute VEMPid, which would facilitate the use of VEMPid in subjects with pathological conditions. A generic template was created by averaging cVEMP responses from 6 normal subjects. To compare VEMPid calculations using a generic versus a subject-specific template, cVEMPs were obtained in 40 healthy subjects using 500, 750, and 1000 Hz tonebursts at sound levels ranging from 98 to 123 dB peSPL. VEMPids were calculated both with the generic template and with the subject's own template. The ability of both templates to determine whether a cVEMP was present or not was compared with receiver operating characteristic curves. No significant differences were found between VEMPid calculations using a generic template versus using a subject-specific template for all frequencies and sound levels. Based on the receiver operating characteristic curves, the subject-specific and generic template did an equally good job at determining threshold. Within limits, the shape of the generic template did not affect these results. A generic template can be used instead of a subject-specific template to calculate VEMPid. Compared with cVEMP normalized by electromyographic amplitudes, VEMPid is advantageous because it averages zero when there is no sound stimulus and it allows the accumulating VEMPid value to be shown during data acquisition as a guide to deciding when enough data has been collected.

Coordinated activities of trunk and upper extremity muscles during walker-assisted paraplegic gait: A synergy study

Individuals with spinal cord injury (SCI) at lower thoracic levels might walk independently with the aid of mechanical orthoses and walker by using their unimpaired trunk and upper extremity muscles (TUEM). The required motor skills and the associated subtasks of the paraplegic locomotion, however, have not been well understood yet. The purpose of this study was to investigate the coordination of the TUEM activities throughout the paraplegic gait cycle using synergy analysis. For eight paraplegic individuals (30.6 ± 11.6 years; SCI level: T12), the kinematics data and the surface electromyography (EMG) signals of TUEM were recorded during 15 gait cycles. Non-negative matrix factorization method was used to extract muscle synergies from the EMG results, which were time normalized in association with four stages of each gait step. For each subject, the number, structures and activation profiles of the muscle synergies across different gait cycles were highly similar, resulting in three subject-specific templates of synergy structure and activation profile with mean intra-subject similarity scores of 0.94 ± 0.02 (p < 0.037) and 0.90 ± 0.06 (p < 0.001), respectively. Hierarchical clustering of the subject-specific templates revealed five representative templates for all subjects, among which, three that were in common for at least six subjects were considered as the characteristic synergies of paraplegic locomotion. The first and second characteristic synergies (inter-subject similarity scores >0.91, p < 0.037) composed mainly of the pectoralis major, latissimus dorsi, longissimus, iliocostalis quadratus lumborum, and the triceps, posterior deltoid and lower trapezius muscles, respectively, and were activated during the ipsilateral and the contralateral leg raising and swing stages. The third characteristic synergy, (inter-subject similarity scores >0.95, p < 0.02), contained mainly the abdominal and lumbopelvic muscles and was activated during the balance adjustment and walker propulsion stages of the gait cycle. These results provide more insight into the motor skills and the associated subtasks of the paraplegic locomotion.

Brain structural plasticity with spaceflight.

Humans undergo extensive sensorimotor adaptation during spaceflight due to altered vestibular inputs and body unloading. No studies have yet evaluated the effects of spaceflight on human brain structure despite the fact that recently reported optic nerve structural changes are hypothesized to occur due to increased intracranial pressure occurring with microgravity. This is the first report on human brain structural changes with spaceflight. We evaluated retrospective longitudinal T2-weighted MRI scans and balance data from 27 astronauts (thirteen ~2-week shuttle crew members and fourteen ~6-month International Space Station crew members) to determine spaceflight effects on brain structure, and whether any pre to postflight brain changes are associated with balance changes. Data were obtained from the NASA Lifetime Surveillance of Astronaut Health. Brain scans were segmented into gray matter maps and normalized into MNI space using a stepwise approach through subject specific templates. Non-parametric permutation testing was used to analyze pre to postflight volumetric gray matter changes. We found extensive volumetric gray matter decreases, including large areas covering the temporal and frontal poles and around the orbits. This effect was larger in International Space Station versus shuttle crew members in some regions. There were bilateral focal gray matter increases within the medial primary somatosensory and motor cortex; i.e., the cerebral areas where the lower limbs are represented. These intriguing findings are observed in a retrospective data set; future prospective studies should probe the underlying mechanisms and behavioral consequences.

Quantitative analysis of glenoid bone loss in osteoarthritis using three-dimensional computed tomography scans

The 3-dimensional (3D) shape of the glenoid vault has been defined previously and shown to be a complex, yet consistent, shape in individuals without glenoid pathology. We proposed assessing whether this conserved shape could be used as a template to measure glenoid bone loss in subjects with glenohumeral osteoarthritis. Computed tomography (CT) scans of both shoulders were obtained from 12 subjects with unilateral glenohumeral osteoarthritis. The paired scapulae were reconstructed 3-dimensionally, using a previously developed graphic software package. Two methods of estimating glenoid bone loss were performed. First, using the software, a stereolithography model of the standardized vault shape was implanted into each glenoid and measurements made of the volume of the implant not contained within each vault. Second, direct measurements of the paired glenoid vault volumes were performed. The volume of the nonarthritic glenoid was used as a subject-specific template for normal glenoid vault volume for each pair. The glenoid bone volumes measured by each method were compared and Pearson's correlation coefficient determined. The average measurement of glenoid bone loss using the vault implant was within 0.8% (SD +/- 1.5%) of the measurement made using the contralateral, normal glenoid. For all patients, Pearson's correlation coefficient was .99, indicating a very high correlation between the two methods of measuring bone loss (P < .0001). The intricate, yet consistent 3D shape of the glenoid vault can be used as an accurate and reliable template to measure glenoid bone loss in glenohumeral osteoarthritis.

Combining Images Across Multiple Subjects

This article introduces a Bayesian hierarchical model for combining information across multiple images. Our work was motivated by an invasive functional brain mapping technique called direct cortical electrical interference that gives a sparse set of binary observations of an underlying “true” region at multiple sites on the brain surface. To model region shapes that may vary widely across individuals, we use mixtures of simple templates, for example, circles. These subject-specific templates are treated as random effects, governed by a set of population templates that make up a population region. The numbers of subject-specific and population templates are treated as unknown variables to be estimated from the data. Conditional on the subject-specific regions, the observed data are modeled using logistic regression. To estimate the variability among images across patients, we develop a measure based on Baddeley's error measure for binary images. Because the dimension of the parameter space changes as the numbers of subject-specific and population templates change, inference is made using reversible jump Markov chain Monte Carlo. Using a hierarchical approach, we may better estimate each individual's region by borrowing strength from other subjects' data, we can estimate a population region by pooling information across subjects, and we can use a collection of data from previous patients to predict the location of a future patient's region of interest. The approach is illustrated with DCEI data collected on 20 patients for two motor tasks: tongue and hand movements.