Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “multimodal data”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 199 records · Page 11Linked to original sources
PubMed2001

Increased jugular bulb saturation is associated with poor outcome in traumatic brain injury.

The objective was to compare secondary insults, particularly decreases in jugular bulb oxyhaemoglobin saturation (SjO(2)), during intensive care in patients with "poor" and "good" outcomes 12 months after traumatic brain injury. A prospective observational study of patients' physiological data collected each minute from multimodality monitoring was carried out. Patients had duration of physiological insults quantified as a percentage of their validated monitoring time (once invalid data due to technical reasons were removed). Treatment protocols were designed to minimise secondary insults by maintaining intracranial pressure (ICP) less than 20 mm Hg, and cerebral perfusion pressure (CPP) greater than 70 mm Hg, with prompt correction of hypoxia and pyrexia. Twelve months after injury patients' neurological function was assessed using the Glasgow outcome scale (GOS). A poor outcome was defined as GOS 1 to 3 (group 1) and a good outcome as GOS 4 and 5 (group 2). Seventy five patients (64 male), median age of 34 years (range 15 to 70), were studied. At 12 months 33 patients had a poor outcome (group 1), and 42 a good outcome (group 2). Group 1 spent proportionately more time with SjO(2) greater than 75% compared with group 2 (p<0.05), and more time with SjO(2) below 54% (p<0.04). Group 1 patients also spent proportionately more time with CPP less than 70 mm Hg than group 2 (p<0.04). Patients in group 1 were older (p<0.04) and had a lower postresuscitation Glasgow coma score (p<0.002). There was no difference between the groups for ICP, injury severity score, peripheral pulse saturation, and pyrexia. This study confirms that secondary insults, including an increased SjO(2), occur significantly more in patients with poor outcomes. More research into strategies to reduce the impact of secondary insults, including management of increased SjO(2), is required.

Adolescent↗
PubMed1994

The National Cancer Data Base. Report on melanoma.

BACKGROUND: Previous Commission on Cancer Studies have examined time trends in stage of disease and treatment patterns for melanoma. Reported herein are the most current National Cancer Data Base (NCDB) data for melanoma. METHODS: Two calls for data have yielded 20,165 melanoma cases for 1985, 1988, and 1990 from hospital cancer registries across the country. RESULTS: There was a marked increase in Stage 0 (in situ) tumors between 1985 and 1990. NCDB data confirm a trend for more conservative surgical excision of skin melanoma. Use of multimodality therapy was infrequent. CONCLUSIONS: Survival data confirmed an excellent correlation between stage of disease and outcome. Early skin melanoma is highly curable.

Adult↗
PubMed1997

[The multimodal therapy of surgical sepsis].

The data of clinical current of sepsis in 239 wounded with syndrome of system inflammatory reaction (SIR) are investigated. Frequency of sepsis in complications of gun-shot wounds has made 5.4%. At the same time syndrome SIR is diagnosed in 23.4% of cases, that was the indication to realization of the whole complex of antiseptic treatment. Therapy of sepsis included some components: drainage and removal of the septic center; the system antibacterial therapy; the controlled hypocoagulation; immunotherapy; active detoxication; power and trophic support on the basis of balanced feed. The used principles of early diagnostics and multicomponent therapy of surgery of sepsis have allowed to lower lethality up to 29.3% and in sepsis of wound--up to 9.1%.

Combined Modality Therapy↗
PubMed2001

Multimodal characterisation of cortical areas by multivariate analyses of receptor binding and connectivity data.

Cortical areas are regarded as fundamental structural and functional units within the information processing networks of the brain. Their properties have been described extensively by cyto-, myelo- and chemoarchitectonics, cortical and extracortical connectivity patterns, receptive field mapping, activation properties, lesion effects, and other structural and functional characteristics. Systematic integrative approaches aiming at multimodal characterisations of cortical areas or at the delineation of global features of the cortical network, however, are still scarce and usually limited to a single data modality, such as cytoarchitectonical or tract tracing data. Here we describe a methodological framework for the systematic evaluation, comparison and integration of different data modalities from the brain and demonstrate its practical application and significance in the analysis of receptor binding and connectivity data within the motor and visual cortices of macaque monkeys. The framework builds on algorithmic methods to convert data between different cortical parcellation schemes, as well as on statistical techniques for the exploration of multivariate data sets comprising data of different types and scales. Thereby, we establish a relationship between intrinsic area properties as expressed by quantitative receptor binding, and extrinsic inter-area communication, which relies on anatomical connectivity. Our analyses provide preliminary evidence for a good correspondence of these two data types in the motor cortex, and their partial discrepancy in the visual cortex, raising hypotheses about the different organisational aspects highlighted by receptors and connectivity. The methodological framework presented here is flexible enough to accommodate a wide range of further data modalities, and is specific enough to permit novel insights and predictions concerning brain organisation. Thus, this approach promises to be very useful in the endeavour to characterise multimodal structure-function relationships in the brain.

Animals↗
PubMed1997

[The robotization of neurosurgery: state of the art and future outlook].

Neurosurgery is by excellence a field of application for robots, based on multimodal image guidance. Specific motorized tools have been already developed and routinely applied in stereotaxy to position a probe holder or in conventional neurosurgery to hold a microscope oriented towards a given target. The potentialities of these approaches have triggered industrial developments currently commercially available. These systems use data bases, primarily coming from multimodal numerical images from X-ray radiology to magnetic resonance imaging. These spatially encoded data are transferred through digital networks to workstations where images can be processed and surgical procedures are preplanned, then transferred to the robotic systems to which they are connected. We have been using a stereotactic robot since 1989 and a microscope robot since 1995 in various surgical routine procedures. The future of these applications mainly rely on the technical progress in informatics, about image recognition to adapt the preplanning to the actual surgical situation, to correct brain shifts for instance, about image fusion, integrated knowledge such such as brain atlases, as well as virtual reality. The future developments, covering surgical procedure, research and teaching, will sure be far beyond our wildest expectations.

Forecasting↗
PubMed2005

Safety and efficacy of intravenous tissue plasminogen activator stroke treatment in the 3- to 6-hour window using multimodal transcranial Doppler/MRI selection protocol.

BACKGROUND: Growing data point toward intravenous tissue plasminogen activator (tPA) benefit after 3 hours in selected stroke patients. We aim to study safety and efficacy of tPA treatment in the 3- to 6-hour window using multimodal transcranial Doppler (TCD)/MRI selection criteria. METHODS: We studied patients with acute middle cerebral artery (MCA) occlusion. Patients within 0 to 3 hours from symptom onset (A) were treated according to standard computed tomography criteria. Treatment within 3 to 6 hours (B) was decided according to TCD/MRI protocol. Continuous TCD assessed clot location and recanalization. National Institutes of Health Stroke Scale (NIHSS) at 24 hours assessed neurological improvement/worsening and modified Rankin score <3 functional independence at third month. RESULTS: Of 135 patients, 56 were in the 3- to 6-hour window. Only 13 (23%) patients within 3 to 6 hours did not meet MRI inclusion criteria. Finally, 122 patients were treated with tPA: A, 79 (65%); B, 43 (35%). Median time to treatment was: A, 136 minutes (range 60 to 180); B, 223 (185 to 360). There were no differences in demographic parameters, baseline NIHSS (A, 17; B, 17; P=0.89), and occlusion location (proximal MCA A, 65.8%; B, 74.4%; P=0.28). Recanalization rates at 2 hours were similar (A, 49.3%; B, 55.2%; P=0.33), as were hemorrhagic transformation rates (asymptomatic: A, 18.7%, B, 26.6%, P=0.43; symptomatic: A, 3.75%, B, 2.38%, P=0.66). Improvement at discharge was similar in both groups (NIHSS dropped 6.3 points [A] versus 6.1 [B]; P=0.86). However, the number of patients who benefited from treatment was slightly higher in the 3- to 6-hour group (A, 58.2%; B, 76.2%; P=0.05), whereas the same rate of patients worsened (A, 11.4%; B, 7.1%; P=0.46). At 3 months, the rate of independent patients was: A, 42% versus B, 38% (P=0.74). CONCLUSIONS: tPA treatment can be safely and effectively extended to the 3- to 6-hour window using TCD/MRI selection criteria. Not using these criteria in the 3- to 6-hour window avoids potentially effective treatment in a high rate of patients.

Acute Disease↗
PubMed2005

Use of ultra-high-resolution data for temporal bone dissection simulation.

OBJECTIVES: For the past 5 years, our group has been developing a virtual temporal bone dissection environment for training otologic surgeons. Throughout the course of our development, a recurring challenge is the acquisition of high-resolution, multimodal, and multi-scale data sets that are used for the visual as well as haptic (sense of touch) display. This study presents several new techniques in temporal bone imaging and their use as data for surgical simulation. METHODS: At our institution (OSU), we are fortunate to have a high-field (8 Tesla) magnetic resonance imaging (MRI) research magnet that provides an order of magnitude higher resolution compared to clinical 1.5T MRI scanners. Magnetic resonance imaging has traditionally been superb at delineating soft tissue structure, and certainly, the 8T unit does indeed do this at a resolution of 100-200 microm(3). To delineate the bony structure of the mastoid and middle ear, computed tomography (CT) has traditionally been used because of the high signal-to-noise ratio delineating bone signal from air and soft tissue. We have partnered with researchers at other institutions (CCF) to make use of a "microCT" that provides a resolution of 214 x 214 x 390 micrometers of bony structure. RESULTS: This report provides a description of the 2 methodologies and presentation of the striking image data capable of being generated. See images presented. CONCLUSIONS: Using these 2 new and innovative imaging modalities, we provide an order of magnitude greater resolution to the visual and haptic display in our temporal bone dissection simulation environment.

Cadaver↗
PubMed1997

Acceptance of a speech interface for biomedical data collection.

Speech interfaces have the potential to address the data entry bottleneck of many applications is the field of medical informatics. An experimental study evaluated the effect of perceptual structure on a multimodal speech interface for the collection of histopathology data. A perceptually structured multimodal interface, using speech and direct manipulation, was shown to increase speed and accuracy. Factors influencing user acceptance are also discussed.

Attitude to Computers↗
PubMed2005

Esthesioneuroblastoma in childhood and adolescence. Better prognosis with multimodal treatment?

BACKGROUND AND PURPOSE: Only 3% of all malignant intranasal tumors are esthesioneuroblastomas (ENB) and only 20% of these rare neuroectodermal tumors are diagnosed up to 20 years of age. Radiotherapy and surgery are established treatment modalities for these patients, but the role of chemotherapy, especially in a multimodal approach, is not well defined. To investigate the influence of radio- and chemotherapy, the treatment and course of the disease in children and adolescents with ENB were analyzed retrospectively. PATIENTS AND METHODS: 19 unselected patients (nine male and ten female) diagnosed with ENB < or = 20 years of age were included in this analysis. Median age at diagnosis was 14.0 years (range, 5-20 years). The tumors were Kadish stage B in 4/19 patients and stage C in 15/19 patients. 17 patients underwent surgery, either without further therapy (n = 4), followed by radiotherapy (n = 1) or as part of multimodal regimens (n = 12). Two patients received radio- and chemotherapy without surgery. Complete resection (R0) was achieved in 15 out of 17 patients with surgery including all five patients with preoperative chemotherapy due to unresectable primary at diagnosis. RESULTS: The 5-year overall survival (OS) for the whole group was 73% +/- 12% and the 5-year event-free survival (EFS) 55% +/- 13%. None of the four patients with stage B experienced tumor progression so far, whereas seven out of 15 patients with stage C did (5-year EFS 47% +/- 14%; not significant). Patients with Kadish stage C and multimodal treatment strategies combing surgery, chemo- and radiotherapy had a significantly better outcome than patients with stage C and less than three treatment modalities (65% +/- 17% vs. 20% +/- 18%; p = 0.02). CONCLUSION: These data indicate a benefit of multimodal treatment regimens combining surgery, chemo- and radiotherapy for pediatric patients with ENB Kadish stage C. Chemotherapy appears to improve resectability, EFS, and OS. Radiotherapy is an integral part in the management of children and young adolescents with ENB in Kadish stage B and C.

Adolescent↗
PubMed2026

NeuroOmics-Net: An interpretable multimodal deep learning framework for Alzheimer's disease diagnosis and progression prediction using neuroimaging, EEG, and genomic data.

Accurate diagnosis and progression prediction of Alzheimer's disease (AD) remain challenging due to the heterogeneous nature of the disease, which involves structural brain degeneration, electrophysiological dysfunction, and molecular dysregulation. Most existing deep learning approaches rely on a single modality or limited multimodal combinations, thereby failing to capture the complex cross-domain interactions underlying AD progression. Furthermore, the scarcity of large-scale datasets containing synchronized neuroimaging, electrophysiological, and genomic measurements restricts the development of comprehensive multimodal diagnostic systems. To address these challenges, this study proposes NeuroOmics-Net, a multimodal deep learning framework for Alzheimer's disease analysis that integrates structural magnetic resonance imaging (sMRI), electroencephalography (EEG), and gene expression data. The proposed framework combines a Hierarchical Multi-View Encoder (HME) for modality-specific feature extraction, a Cross-Omics Attention Fusion (CAF) module for adaptive integration of complementary biomarkers, and a Disease Progression Graph Learning (DPGL) module for modeling progression-related relationships across biological domains. To facilitate cross-modal integration from independent cohorts, Regularized Canonical Correlation Analysis (RCCA) is employed to align heterogeneous feature representations within a shared latent space. Experiments were conducted using publicly available datasets from ADNI, PhysioNet, and GEO repositories comprising 1120 diagnosis-aligned samples. The proposed framework achieved 94.3% classification accuracy and an AUC of 0.975 for distinguishing normal controls (NC), mild cognitive impairment (MCI), and Alzheimer's disease subjects, while attaining 93.7% accuracy for predicting conversion from stable mild cognitive impairment (sMCI) to progressive mild cognitive impairment (pMCI). However, a fairness sensitivity analysis using stratified demographic reweighting revealed accuracy ranging from 90.8% (low-education, high-comorbidity proxy subgroup) to 96.1% (low-risk, high-reserve proxy subgroup), a demographic parity gap of 5.3 percentage points, indicating that overall accuracy reflects a performance ceiling in a relatively homogeneous research cohort rather than a realistic estimate for demographically diverse clinical populations. Comparative evaluations demonstrated consistent improvements over state-of-the-art unimodal and multimodal deep learning models. Interpretability analysis further identified clinically relevant biomarkers, including hippocampal and entorhinal atrophy, theta-alpha EEG alterations, and APOE-associated molecular pathways. Because sMRI, EEG, and gene expression data were sourced from separate, unpaired cohorts with no subjects possessing all three synchronized measurements, all reported cross-modal associations reflect population-level statistical correspondence across diagnosis-matched groups rather than within-subject physiological coupling; no claim of intra-individual causal cross-modal interaction is made. These findings demonstrate that NeuroOmics-Net provides an effective computer-aided framework for multimodal biomedical data processing and Alzheimer's disease analysis. By integrating neuroimaging, electrophysiological, and genomic information, the proposed approach enables accurate diagnosis, progression prediction, and biologically interpretable decision support for clinical and translational applications.

Humans↗
PubMed2004

Vacuum ultraviolet mass-analyzed threshold ionization spectroscopy of hexafluorobenzene: the Jahn-Teller effect and vibrational analysis.

One-photon mass-analyzed threshold ionization (MATI) spectrum of hexafluorobenzene was obtained by using vacuum ultraviolet radiation generated by four-wave difference frequency mixing in Kr. The ionization energy of hexafluorobenzene determined from the position of the 0-0 band was 9.9108+/-0.0006 eV. To aid the spectral analysis, the Jahn-Teller coupling parameters for four e(2g) modes of C(6)F(6) (+) in the ground electronic state were calculated from the topographical data of the potential energy surface obtained at the density functional theory (DFT) level. These were used in the initial calculation of the energies of the Jahn-Teller states and upgraded through the multimode fit to the experimental data. Excellent agreement between the experimental and calculated frequencies was achieved. The vibrations which are not linear Jahn-Teller active were observed and could be assigned by referring to the frequencies obtained at the DFT level.

Journal Article↗
PubMed2003

AMIDE: a free software tool for multimodality medical image analysis.

Amide's a Medical Image Data Examiner (AMIDE) has been developed as a user-friendly, open-source software tool for displaying and analyzing multimodality volumetric medical images. Central to the package's abilities to simultaneously display multiple data sets (e.g., PET, CT, MRI) and regions of interest is the on-demand data reslicing implemented within the program. Data sets can be freely shifted, rotated, viewed, and analyzed with the program automatically handling interpolation as needed from the original data. Validation has been performed by comparing the output of AMIDE with that of several existing software packages. AMIDE runs on UNIX, Macintosh OS X, and Microsoft Windows platforms, and it is freely available with source code under the terms of the GNU General Public License.

Costs and Cost Analysis↗
PubMed2006

Multiscale hybrid linear models for lossy image representation.

In this paper, we introduce a simple and efficient representation for natural images. We view an image (in either the spatial domain or the wavelet domain) as a collection of vectors in a high-dimensional space. We then fit a piece-wise linear model (i.e., a union of affine subspaces) to the vectors at each downsampling scale. We call this a multiscale hybrid linear model for the image. The model can be effectively estimated via a new algebraic method known as generalized principal component analysis (GPCA). The hybrid and hierarchical structure of this model allows us to effectively extract and exploit multimodal correlations among the imagery data at different scales. It conceptually and computationally remedies limitations of many existing image representation methods that are based on either a fixed linear transformation (e.g., DCT, wavelets), or an adaptive uni-modal linear transformation (e.g., PCA), or a multimodal model that uses only cluster means (e.g., VQ). We will justify both quantitatively and experimentally why and how such a simple multiscale hybrid model is able to reduce simultaneously the model complexity and computational cost. Despite a small overhead of the model, our careful and extensive experimental results show that this new model gives more compact representations for a wide variety of natural images under a wide range of signal-to-noise ratios than many existing methods, including wavelets. We also briefly address how the same (hybrid linear) modeling paradigm can be extended to be potentially useful for other applications, such as image segmentation.

Algorithms↗
PubMedDate not supplied

A 3D medical image database management system.

We describe the design and implementation of QBISM (Query By Interactive, Spatial Multimedia), a prototype for querying and visualizing 3D spatial data. Our medical image application is focused on the brain mapping requirements for multimodality relationships across multiple subjects. It incorporates data describing both structure and function. It includes data structures that describe anatomy, physiology, coordinates using rendered imagery and statistical output. The system is built on top of the Starburst DBMS extended to handle spatial data types, specially, scalar fields and arbitrary regions or space within such fields. In this paper we list the requirements of the application, discuss the logical and physical database design issues, and present timing results from our prototype. We observed that the DBMS' early spatial filtering results in significant performance savings because the system response time is dominated by the amount of data retrieved, transmitted, and rendered.

Brain Mapping↗
PubMed2025

Diagnosing the undiagnosed: AI-enhanced multimodal modeling for placental mesenchymal dysplasia in high-risk pregnancies.

Placental mesenchymal dysplasia (PMD) is a rare vascular placental disorder that mimics molar pregnancy but often coexists with a viable fetus, making its misdiagnosis potentially devastating. In high-risk pregnancies, artificial intelligence (AI)-enhanced multimodal modeling - incorporating imaging, genomics, proteomics, and clinical features - offers a transformative diagnostic strategy. Leveraging Bayesian hyperparameter optimization for model refinement, this approach improves diagnostic accuracy while reducing uncertainty and clinician hesitation. Recent clinical studies support its efficacy and interpretability through SHAP and LIME models, while real-time surgical enhancements using Bayesian methods highlight its broader clinical utility. Despite current challenges such as data heterogeneity and integration barriers, multimodal AI provides unprecedented resolution in placental analysis, enabling precise differentiation between PMD and similar fetopathies. Ultimately, this advancement supports timely, non-invasive diagnosis, personalized management, and emotionally informed decision-making aligned with ethical AI implementation standards.

Bayesian optimization↗
PubMed1999

A new class of auditory warning signals for complex systems: auditory icons.

This simulator-based study examined conventional auditory warnings (tonal, nonverbal sounds) and auditory icons (representational, nonverbal sounds), alone and in combination with a dash-mounted visual display, to present information about impending collision situations to commercial motor vehicle operators. Brake response times were measured for impending front-to-rear collision scenarios under 6 display configurations, 2 vehicle speeds, and 2 levels of headway. Accident occurrence was measured for impending side collision scenarios under 2 vehicle speeds, 2 levels of visual workload, 2 auditory displays, absence/presence of mirrors, and absence/presence of a dash-mounted iconic visual display. For both front-to-rear and side collision scenarios, auditory icons elicited significantly improved driver performance over conventional auditory warnings. Driver performance improved when collision warning information was presented through multiple modalities. Brake response times were significantly faster for impending front-to-rear collision scenarios using the longer headway condition. The presence of mirrors significantly reduced the number of accidents for impending side collision scenarios. Subjective preference data indicated that participants preferred multimodal displays over single-modality displays. Actual or potential applications for this research include auditory displays and warnings, information presentation, and the development of alternative user interfaces.

Accidents, Traffic↗
PubMed2002

Long-term results with multimodal adjuvant therapy and liver transplantation for the treatment of hepatocellular carcinomas larger than 5 centimeters.

OBJECTIVE: To determine the long-term results of liver transplantation for hepatocellular carcinoma (HCC) measuring 5 cm or larger treated in a multimodality adjuvant protocol. SUMMARY BACKGROUND DATA: Transplant has been established as a viable treatment of HCC measuring less than 5 cm, but the results for larger tumors have been disappointing. Several studies have shown promising preliminary results when combining transplant with preoperative transarterial chemoembolization and/or perioperative systemic chemotherapy in the treatment of advanced HCC that is not amenable to resection. However, follow-up in the studies has been limited and the number of patients has been small. METHODS: Beginning in October 1991, all patients with unresectable HCC measuring 5 cm or larger, as measured by computed tomography, were considered for enrollment in the authors' multimodality protocol. Entry criteria required that all patients be free of extrahepatic disease based on computed tomography scans of the chest and abdomen and bone scan and have a patent main portal vein and major hepatic veins on duplex ultrasonography. Patients received subselective arterial chemoembolization with mitomycin C, doxorubicin, and cisplatin at the time of diagnosis, repeated as necessary based on tumor response. Patients received a single systemic intraoperative dose of doxorubicin (10 mg/m(2)) before revascularization of the new liver and systemic doxorubicin (50 mg/m(2)) every 3 weeks as tolerated, for a total of six cycles, beginning on the sixth postoperative week. RESULTS: Eighty patients were enrolled; 37 were eventually excluded, due mainly to disease progression while on the waiting list, and 43 underwent liver transplant. Mean pathologic tumor diameter was 5.8 +/- 2.7 cm. Median follow-up of surviving transplanted patients was 55.1 +/- 24.9 months. There were two (4.7%) perioperative deaths. Median overall survival was significantly longer in transplanted patients (49.9 +/- 10.42 months) than in those who were excluded (6.83 +/- 1.34 months). Overall and recurrence-free survival rates in transplanted patients at 5 years were 44% and 48%, respectively. A tumor size larger than 7 cm and the presence of vascular invasion correlated significantly with recurrence. Recurrence-free survival at 5 years was significantly higher for the 32 patients with tumors measuring 5 to 7 cm (55%) than the 12 patients with tumors larger than 7 cm (34%). CONCLUSIONS: A significant proportion of patients with HCC measuring 5 cm or larger can achieve long-term survival after liver transplantation in the context of multimodal adjuvant therapy. Patients with tumors measuring 5 to 7 cm have significantly longer recurrence-free survival compared with those with larger tumors.

Aged↗