Biology subjects

Palombo, M.

Publications and source records attributed to Palombo, M..

25 records · Page 2Linked to original sources

bioRxiv2021

Training Data Distribution Significantly Impacts the Estimation of Tissue Microstructure with Machine Learning

PurposeSupervised machine learning (ML) provides a compelling alternative to traditional model fitting for parameter mapping in quantitative MRI. The aim of this work is to demonstrate and quantify the effect of different training strategies on the accuracy and precision of parameter estimates when supervised ML is used for fitting. MethodsWe fit a two-compartment biophysical model to diffusion measurements from in-vivo human brain, as well as simulated diffusion data, using both traditional model fitting and supervised ML. For supervised ML, we train several artificial neural networks, as well as random forest regressors, on different distributions of ground truth parameters. We compare the accuracy and precision of parameter estimates obtained from the different estimation approaches using synthetic test data. ResultsWhen the distribution of parameter combinations in the training set matches those observed in similar data sets, we observe high precision, but inaccurate estimates for atypical parameter combinations. In contrast, when training data is sampled uniformly from the entire plausible parameter space, estimates tend to be more accurate for atypical parameter combinations but may have lower precision for typical parameter combinations. ConclusionThis work highlights the need to consider the choice of training data when deploying supervised ML for estimating microstructural metrics, as performance depends strongly on the training-set distribution. We show that high precision obtained using ML may mask strong bias, and visual assessment of the parameter maps is not sufficient for evaluating the quality of the estimates.

neuroscience↗

bioRxiv2021

Neural Networks for parameter estimation in microstructural MRI: a study with a high-dimensional diffusion-relaxation model of white matter microstructure

Specific features of white-matter microstructure can be investigated by using biophysical models to interpret relaxation-diffusion MRI brain data. Although more intricate models have the potential to reveal more details of the tissue, they also incur time-consuming parameter estimation that may con-verge to inaccurate solutions due to a prevalence of local minima in a degenerate fitting landscape. Machine-learning fitting algorithms have been proposed to accelerate the parameter estimation and increase the robustness of the attained estimates. So far, learning-based fitting approaches have been restricted to lower-dimensional microstructural models where dense sets of training data are easy to generate. Moreover, the degree to which machine learning can alleviate the degeneracy problem is poorly understood. For conventional least-squares solvers, it has been shown that degeneracy can be avoided by acquisition with optimized relaxation-diffusion-correlation protocols that include tensor-valued diffusion encoding; whether machine-learning techniques can offset these acquisition require-ments remains to be tested. In this work, we employ deep neural networks to vastly accelerate the fitting of a recently introduced high-dimensional relaxation-diffusion model of tissue microstructure. We also develop strategies for assessing the accuracy and sensitivity of function fitting networks and use those strategies to explore the impact of acquisition protocol design on the performance of the network. The developed learning-based fitting pipelines were tested on relaxation-diffusion data acquired with optimized and sub-sampled acquisition protocols. We found no evidence that machine-learning algorithms can by themselves replace a careful design of the acquisition protocol or correct for a degenerate fitting landscape.

neuroscience↗

bioRxiv2021

On the generalizability of diffusion MRI signal representations across acquisition parameters, sequences and tissue types: chronicles of the MEMENTO challenge.

Diffusion MRI (dMRI) has become an invaluable tool to assess the microstructural organization of brain tissue. Depending on the specific acquisition settings, the dMRI signal encodes specific properties of the underlying diffusion process. In the last two decades, several signal representations have been proposed to fit the dMRI signal and decode such properties. Most methods, however, are tested and developed on a limited amount of data, and their applicability to other acquisition schemes remains unknown. With this work, we aimed to shed light on the generalizability of existing dMRI signal representations to different diffusion encoding parameters and brain tissue types. To this end, we organized a community challenge - named MEMENTO, making available the same datasets for fair comparisons across algorithms and techniques. We considered two state-of-the-art diffusion datasets, including single-diffusion-encoding (SDE) spin-echo data from a human brain with over 3820 unique diffusion weightings (the MASSIVE dataset), and double (oscillating) diffusion encoding data (DDE/DODE) of a mouse brain including over 2520 unique data points. A subset of the data sampled in 5 different voxels was openly distributed, and the challenge participants were asked to predict the remaining part of the data. After one year, eight participant teams submitted a total of 80 signal fits. For each submission, we evaluated the mean squared error, the variance of the prediction error and the Bayesian information criteria. Most predictions predicted either multi-shell SDE data (37%) or DODE data (22%), followed by cartesian SDE data (19%) and DDE (18%). Most submissions predicted the signals measured with SDE remarkably well, with the exception of low and very strong diffusion weightings. The prediction of DDE and DODE data seemed more challenging, likely because none of the submissions explicitly accounted for diffusion time and frequency. Next to the choice of the model, decisions on fit procedure and hyperparameters play a major role in the prediction performance, highlighting the importance of optimizing and reporting such choices. This work is a community effort to highlight strength and limitations of the field at representing dMRI acquired with trending encoding schemes, gaining insights into how different models generalize to different tissue types and fiber configurations over a large range of diffusion encodings.

neuroscience↗

bioRxiv2020

SPHERIOUSLY? The challenges of estimating spherical pore size non-invasively in the human brain from diffusion MRI

The Soma and Neurite Density Imaging (SANDI) three-compartment model was recently proposed to disentangle cylindrical and spherical geometries, attributed to neurite and soma compartments, respectively, in brain tissue. The approach could also enable estimation of microstructure parameters such as the apparent size (radius) of the soma. There are some recent advances in diffusion-weighted MRI signal encoding and analysis (including the use of multiple so-called b-tensor encodings and analysing the signal in the frequency-domain) that have not yet been applied in the context of SANDI. In this work, using: (i) ultra-strong gradients; (ii) a combination of linear, planar, and spherical b-tensor encodings; and (iii) analysing the signal in the frequency domain, three main challenges to robust estimation of soma size were identified: First, the Rician noise floor in magnitude-reconstructed data biases estimates of soma properties in a non-uniform fashion. It may cause overestimation or underestimation of the soma size and density. This can be partly ameliorated by accounting for the noise floor in the estimation routine. Second, even when using the strongest diffusion-encoding gradient strengths available for human MRI, there is an empirical lower bound on the spherical signal fraction and pore-size that can be detected and estimated robustly. For the experimental setup used here, the lower bound on the signal fraction was approximately 10%. We employed two different ways of establishing the lower bound for spherical radius estimates in white matter. The first, examining power-law relationships between the DW-signal and diffusion weighting in empirical data, yielded a lower bound of 7 m, while the second, pure Monte Carlo simulations, yielded a lower limit of 3 m and in this low radii domain, there is little differentiation in signal attenuation. Third, if there is sensitivity to the transverse intra-cellular diffusivity in cylindrical structures, e.g., axons and cellular projections, then trying to disentangle two diffusion-time-dependencies using one experimental parameter (i.e., change in frequency-content of the encoding waveform) makes spherical pore-size estimates particularly challenging. We conclude that due to the aforementioned challenges spherical pore size estimates may be biased when the corresponding signal fraction is low, which must be considered when using them as biomarkers in clinical/research studies.

neuroscience↗

bioRxiv2020

Deep learning model fitting for diffusion-relaxometry: a comparative study

Quantitative Magnetic Resonance Imaging (qMRI) signal model fitting is traditionally performed via non-linear least square (NLLS) estimation. NLLS is slow and its performance can be affected by the presence of different local minima in the fitting objective function. Recently, machine learning techniques, including deep neural networks (DNNs), have been proposed as robust alternatives to NLLS. Here we present a deep learning implementation of qMRI model fitting, which uses DNNs to perform the inversion of the forward signal model. We compare two DNN training strategies, based on two alternative definitions of the loss function, since at present it is not known which definition leads to the most accurate, precise and robust parameter estimation. In strategy 1 we define the loss as the l2-norm of tissue parameter prediction errors, while in strategy 2 as the l2-norm of MRI signal prediction errors. We compare the two approaches on synthetic and 3T in vivo saturation inversion recovery (SIR) diffusion-weighted (DW) MRI data, using a model for joint diffusion-T1 mapping. Strategy 1 leads to lower tissue parameter root mean squared errors (RMSEs) when realistic noise distributions are considered (e.g. Rician vs Gaussian). However, strategy 2 offers lower signal reconstruction RMSE, and allows training to be performed on both synthetic and actual in vivo MRI measurements. In conclusion, both strategies are valid choices for DNN-based fitting. Strategy 2 is more practical, as it does not require pre-computation of reference tissue parameters, but may lead to worse parameter estimation.

neuroscience↗

bioRxiv2020

Repeatability of soma and neurite metrics incortical and subcortical grey matter

Diffusion magnetic resonance imaging is a technique which has long been used to study white matter microstructure in vivo. Recent advancements in hardware and modelling techniques have opened up interest in disentangling tissue compartments in the grey matter. In this study, we evaluate the repeatability of soma and neurite density imaging in a sample of six healthy adults scanned five times on an ultra-strong gradient magnetic resonance scanner (300 mT/m). Repeatability was expressed as an intraclass correlation coefficient (ICC). Our findings reveal that measures of soma density (mean ICC=.976), neurite density (mean ICC=.959) and apparent soma size (mean ICC=.923) are highly reliable across multiple cortical and subcortical networks. Overall, we demonstrate the promise of moving advanced grey matter microstructural imaging towards applications of development, ageing, and disease.

neuroscience↗

bioRxiv2018

Diffusion-weighted magnetic resonance spectroscopy enables cell-specific monitoring of astrocyte reactivity in vivo

The diffusion of brain intracellular metabolites, as measured using diffusion-weighted magnetic resonance spectroscopy in vivo, is thought to specifically depend on the cellular structure constraining them. However, it has never been established that variations of metabolite diffusion, e.g. as observed in some diseases, could indeed be linked to alterations of cellular morphology. Here we demonstrate, in a mouse model of reactive astrocytes, that advanced diffusion-weighted magnetic resonance spectroscopy acquisition and modeling techniques enable non-invasive detection of reactive astrocyte hypertrophy (increased soma radius, increased fiber radius and length), as inferred from variations of myo-inositol diffusion, and as confirmed by confocal microscopy ex vivo. This establishes that specific alterations of intracellular metabolite diffusion can be measured and related to cell-specific morphological alterations. Furthermore, as reactive astrocytes are a hallmark of many brain pathologies, this work opens exciting perspectives for neuroscience and clinical research.

bioinformatics↗