Meta-analysis and open-source database for in vivo brain Magnetic Resonance spectroscopy in health and disease

Gudmundson, Aaron T; Koo, Annie; Virovka, Anna; Amirault, Alyssa L.; Soo, Madelene; Cho, Jocelyn H.; Oeltzschner, Georg; Edden, Richard A.E.; Stark, Shauna M.

doi:10.1016/j.ab.2023.115227

Cited by 4 publications

(4 citation statements)

References 374 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Individual metabolite basis functions were linearly combined to give a metabolite spectral component, weighted by metabolite concentrations sampled from distributions defined by our recent meta-analysis (Gudmundson et al, 2023), including both healthy and clinical cohort ranges. From the full basis sets, 22 metabolites were selected which had defined concentration ranges available in a recent meta-analysis that collated results from nearly 500 MRS papers using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (Gudmundson et al, 2023; Moher et al, 2009; Page et al, 2021). One isomer of GABA (either the definition from (Govindaraju et al, 2000) or (Near et al, 2012)) and Glucose (α or β) were randomly chosen with equal probability for each example.…”

Section: Methodsmentioning

confidence: 99%

“…One isomer of GABA (either the definition from (Govindaraju et al, 2000) or (Near et al, 2012)) and Glucose (α or β) were randomly chosen with equal probability for each example. Concentrations were selected with equal probability from a range defined by ±2.5 standard deviations from the meta-analysis mean of each cohort (Gudmundson et al, 2023) and are provided in supplemental tables 1 and 2.…”

Section: Methodsmentioning

confidence: 99%

“…While pure T 2 is understood to be field-independent (Bloembergen et al, 1948; Carr & Purcell, 1954; Held et al, 1973; Michaeli et al, 2002), the dominant Gaussian decay (Marshall et al, 1997) increases with increasing static field strength and is attributed to greater microscopic (Michaeli et al, 2002) and macroscopic (Juchem et al, 2021; Tkáč et al, 2001) susceptibility gradients. Here, the pure Lorentzian T 2 component is based upon the relaxation times at 1.5 T from a relaxation meta-regression (Gudmundson et al, 2023), which are assumed to be the least impacted by susceptibility gradients that scale with B 0 (Bloembergen et al, 1948; De Graaf et al, 2006; Michaeli et al, 2002). Once the Lorentzian T 2 component was applied, the additional T 2 * contributions were modeled by applying appropriate amounts of Gaussian broadening, to achieve a frequency-domain full-width half-maximum ( FWHM ) linewidth of the NAA singlet between 3 Hz and 18 Hz with a uniform distribution.…”

Section: Methodsmentioning

confidence: 99%

See 2 more Smart Citations

Application of a¹H Brain MRS Benchmark Dataset to Deep Learning for Out-of-Voxel Artifacts

Gudmundson

Davies-Jenkins

Özdemir

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

PurposeNeural networks are potentially valuable for many challenges associated with MRS data. The purpose of this manuscript is to describe the AGNOSTIC dataset, which contains 259,200 synthetic MRS examples. To demonstrate the utility, we use AGNOSTIC to train two Convolutional Neural Networks (CNNs) to address out-of-voxel (OOV) echoes. A Detection Network was trained to identify the point-wise presence of OOV echoes, providing proof of concept for real-time detection. A Prediction Network was trained to reconstruct OOV echoes, allowing subtraction during post-processing.MethodsAGNOSTIC was created using 270 basis sets that were simulated across 18 field strengths and 15 echo times. The synthetic examples were produced to resemblein vivobrain data with combinations of metabolite, macromolecule, and residual water signals, and noise.Complex OOV signals were mixed into 85% of synthetic examples to train two separate U-net CNNs for the detection and prediction of OOV signals.ResultsAGNOSTIC is available through Dryad and all Python 3 code is available through GitHub. The Detection network was shown to perform well, identifying 95% of OOV echoes. Traditional modeling of these detected OOV signals was evaluated and may prove to be an effective method during linear-combination modeling. The Prediction Network greatly reduces OOV echoes within FIDs and achieved a median log10normed-MSE of –1.79, an improvement of almost two orders of magnitude.ConclusionThe AGNOSTIC benchmark dataset for MRS is introduced and various dataset features are described. As an exemplar use of AGNOSTIC, two CNNs were developed to detect and predict OOV echoes.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Application of a¹H Brain MRS Benchmark Dataset to Deep Learning for Out-of-Voxel Artifacts

Gudmundson

Davies-Jenkins

Özdemir

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…They are not limited to the aforementioned example areas, but include all but the most basic acquisition paradigms, for example: metabolic kinetic measurements in phosphorus MRS, 13 deuterium metabolic imaging, 14 and time-resolved carbon imaging (both hyperpolarised 15 and non-hyperpolarised 16 ); "Fingerprinting" techniques which aim to elucidate multiple metabolic state parameters 17 ; and multi-inversion and multi-echo techniques that measure relaxation parameters, 18 which might be sensitive markers of pathology. 19,20 There may also be ample application for the better processing of even straightforward acquisition paradigms. 1 Dynamic fitting has repeatedly been demonstrated as advantageous over independent "1D" fitting for specific use cases: initially for T 1 determination by inversion recovery, 21 and then incorporating T 2 measurements as well, 21,22 and even with extensions for determining unknown macromolecule components 18,23,24 ; in fitting 2D j-resolved spectroscopy 25,26 ; in fitting time resolved x-nuclear data 27 ; fitting edited data 28 ; and diffusion, 29 including more complex, non-Gaussian diffusion models 30 ; and even extends to multi-voxel MRSI.…”

Section: Introductionmentioning

confidence: 99%

Universal dynamic fitting of magnetic resonance spectroscopy

Clarke,

Ligneul,

Cottaar

et al. 2024

Magnetic Resonance in Med

View full text Add to dashboard Cite

PurposeDynamic (2D) MRS is a collection of techniques where acquisitions of spectra are repeated under varying experimental or physiological conditions. Dynamic MRS comprises a rich set of contrasts, including diffusion‐weighted, relaxation‐weighted, functional, edited, or hyperpolarized spectroscopy, leading to quantitative insights into multiple physiological or microstructural processes. Conventional approaches to dynamic MRS analysis ignore the shared information between spectra, and instead proceed by independently fitting noisy individual spectra before modeling temporal changes in the parameters. Here, we propose a universal dynamic MRS toolbox which allows simultaneous fitting of dynamic spectra of arbitrary type.MethodsA simple user‐interface allows information to be shared and precisely modeled across spectra to make inferences on both spectral and dynamic processes. We demonstrate and thoroughly evaluate our approach in three types of dynamic MRS techniques. Simulations of functional and edited MRS are used to demonstrate the advantages of dynamic fitting.ResultsAnalysis of synthetic functional 1H‐MRS data shows a marked decrease in parameter uncertainty as predicted by prior work. Analysis with our tool replicates the results of two previously published studies using the original in vivo functional and diffusion‐weighted data. Finally, joint spectral fitting with diffusion orientation models is demonstrated in synthetic data.ConclusionA toolbox for generalized and universal fitting of dynamic, interrelated MR spectra has been released and validated. The toolbox is shared as a fully open‐source software with comprehensive documentation, example data, and tutorials.

show abstract

Glutamatergic basis of antipsychotic response in first-episode psychosis: a dual voxel study of the anterior cingulate cortex

Fan,

Liang,

Wang

et al. 2023

Neuropsychopharmacol.

View full text Add to dashboard Cite

Treatment-resistant schizophrenia is believed to be related to excess prefrontal glutamate. If we can identify these individuals early in the course of illness, the repeated use of rst-line antipsychotics can be reduced and rst-episode patients can be strati ed for accelerated treatments. The use of proton magnetic resonance spectroscopy (1H-MRS) to measure glutamate and Glx (glutamate plus glutamine) provides a means for such a strati cation, but we must rst establish if there is robust evidence linking elevations in anterior cingulate cortex (ACC) glutamate metabolites to poor response, and determine if continuous exposure to antipsychotics worsens the glutamatergic excess in eventual non-responders. In this study, we estimated glutamate levels at baseline in 70 drug-naive patients with schizophrenia. We then treated them (N=42) with risperidone and followed them up for 3 months to categorize their response status. We hypothesized to see "hyperglutamatergia" at baseline in later non-responders and expected this to worsen with treatment. Non-responders had high glutamate before treatment-onset (F1,79=3.20, p=0.046, partial η2 = 0.075); However, glutamate levels did not change signi cantly over time in both non-responders and responders over the 3 months of treatment (F1,31=1.26, p=0.270, partial η2 = 0.039). Antipsychotic use without prior knowledge of later response delays symptom relief in a subgroup of rst-episode patients, but does not worsen the glutamatergic excess seen at baseline. Given the current practice of non-strati ed use of antipsychotics, longer-time follow-up MRS studies are required to see if improvement in symptoms accompanies a shift in glutamate pro le.

show abstract

Meta-analysis and open-source database for in vivo brain Magnetic Resonance spectroscopy in health and disease

Cited by 4 publications

References 374 publications

Application of a¹H Brain MRS Benchmark Dataset to Deep Learning for Out-of-Voxel Artifacts

Application of a¹H Brain MRS Benchmark Dataset to Deep Learning for Out-of-Voxel Artifacts

Universal dynamic fitting of magnetic resonance spectroscopy

Glutamatergic basis of antipsychotic response in first-episode psychosis: a dual voxel study of the anterior cingulate cortex

Contact Info

Product

Resources

About

Meta-analysis and open-source database for in vivo brain Magnetic Resonance spectroscopy in health and disease

Cited by 4 publications

References 374 publications

Application of a1H Brain MRS Benchmark Dataset to Deep Learning for Out-of-Voxel Artifacts

Application of a1H Brain MRS Benchmark Dataset to Deep Learning for Out-of-Voxel Artifacts

Universal dynamic fitting of magnetic resonance spectroscopy

Glutamatergic basis of antipsychotic response in first-episode psychosis: a dual voxel study of the anterior cingulate cortex

Contact Info

Product

Resources

About

Application of a¹H Brain MRS Benchmark Dataset to Deep Learning for Out-of-Voxel Artifacts

Application of a¹H Brain MRS Benchmark Dataset to Deep Learning for Out-of-Voxel Artifacts