Dr. Nazari Farsani works on developing and implementing machine learning techniques for automated tissue segmentation from brain PET/MR images. She is also developing machine learning algorithms for PET data correction and de-noising.
Honors & Awards
Learning to Detect and Segment Brain Lesions with Tailored Context-Aware Deep Neural Networks, Finnish Academy of Science and Letters (2021)
Radiostereometric analysis of femoral stem migrations, Turku University Hospital State Funding (2020)
Imaging of the stability of uncemented femoral stems after hip replacement, Instrumentarium Foundation (2017)
Imaging of the stability of uncemented femoral stems after hip replacement, Finnish Cultural Foundation (2017)
Radiostereometric analysis of femoral stem migrations, University of Turku Foundation (2017)
Gender issue in total hip arthroplasty, Suomen Artroplastiayhdistys (2016)
Neuronavigation in rTMS treatment of tinnitus, Tekniikan edistämissäätio (Finnish Foundation for Technology Promotion) (2015)
Master of Science, Abo Akademi (2020)
Doctor of Philosophy, University Of Turku (2020)
Ph.D., University of Turku, Biomedical Imaging (2020)
M.Sc., Åbo Akademi University, Biomedical Imaging (2014)
B.Sc., Azad University, Biomedical Engineering, Bioelectric (2012)
Craig Levin, Postdoctoral Faculty Sponsor
Decoding Music-Evoked Emotions in the Auditory and Motor Cortex.
Cerebral cortex (New York, N.Y. : 1991)
2021; 31 (5): 2549-2560
Music can induce strong subjective experience of emotions, but it is debated whether these responses engage the same neural circuits as emotions elicited by biologically significant events. We examined the functional neural basis of music-induced emotions in a large sample (n = 102) of subjects who listened to emotionally engaging (happy, sad, fearful, and tender) pieces of instrumental music while their hemodynamic brain activity was measured with functional magnetic resonance imaging (fMRI). Ratings of the four categorical emotions and liking were used to predict hemodynamic responses in general linear model (GLM) analysis of the fMRI data. Multivariate pattern analysis (MVPA) was used to reveal discrete neural signatures of the four categories of music-induced emotions. To map neural circuits governing non-musical emotions, the subjects were scanned while viewing short emotionally evocative film clips. The GLM revealed that most emotions were associated with activity in the auditory, somatosensory, and motor cortices, cingulate gyrus, insula, and precuneus. Fear and liking also engaged the amygdala. In contrast, the film clips strongly activated limbic and cortical regions implicated in emotional processing. MVPA revealed that activity in the auditory cortex and primary motor cortices reliably discriminated the emotion categories. Our results indicate that different music-induced basic emotions have distinct representations in regions supporting auditory processing, motor control, and interoception but do not strongly rely on limbic and medial prefrontal regions critical for emotions with survival value.
View details for DOI 10.1093/cercor/bhaa373
View details for PubMedID 33367590
Contributing factors to the initial femoral stem migration in cementless total hip arthroplasty of postmenopausal women
JOURNAL OF BIOMECHANICS
2021; 117: 110262
In cementless total hip arthroplasty (THA), femoral stems rely on the initial press-fit fixation against cortical bone to achieve osseointegration. Decreased bone mineral density (BMD) in postmenopausal women poses natural difficulties in achieving axial and rotational femoral stem stability. The present study examined contributing demographic, surgery-related and postoperative factors in determining the magnitude of early stem migration prior to osseointegration. A prospective cohort of 65 postmenopausal women with hip osteoarthritis (Dorr type A or B femur anatomy) underwent THA with implantation of an uncemented parallel-sided femoral component. Postoperative femoral stem translation and rotation were measured using model-based radiostereometric analysis. Based on analysis of covariance, which controlled for outliers and randomized antiresorptive treatment with denosumab or placebo, none of the analyzed demographics (including BMI) and surgery-related variables (including the stem-to-canal fil ratio) was associated with stem subsidence. Stem subsidence (mean 1.8 mm, 95% CI 1.2 to 2.4) occurred even in women with normal hip BMD. Total hip BMD and postoperative walking activity (measured three months after surgery) were significantly associated with stem rotation, and height acted as a confounding factor. The effect of walking activity on stem rotation was significant at 5 months (p = 0.0083) and at 11 months (p = 0.0117). This observation confirms the previous results of instrumented hip prostheses on torsional moments affecting stems during daily activities. High-resolution imaging modalities of local bone quality are needed to explore reasons for RSA-measurable stem subsidence even in women with normal hip BMD.
View details for DOI 10.1016/j.jbiomech.2021.110262
View details for Web of Science ID 000626266300006
View details for PubMedID 33508723
- Denosumab in Prevention of Implant Migration JOURNAL OF BONE AND MINERAL RESEARCH 2020; 35 (9): 1824–25
Bone Mineral Density and Cortical-Bone Thickness of the Distal Radius Predict Femoral Stem Subsidence in Postmenopausal Women
JOURNAL OF ARTHROPLASTY
2020; 35 (7): 1877-+
The distal radius is an optional site for evaluation of bone quality in postmenopausal women before cementless total hip arthroplasty. We hypothesized that dual-energy X-ray absorptiometry (DXA) and pulse-echo ultrasonometry of the distal radius may help discriminate subjects at high risk of femoral stem subsidence.A prospective cohort of postmenopausal women with primary hip osteoarthritis underwent total hip arthroplasty with implantation of a parallel-sided femoral stem. Postoperative stem migration was measured using radiostereometric analysis. Preoperatively, subjects had multisite DXA measurement of bone mineral density (BMD) and pulse-echo ultrasonometry of the cortical-bone thickness. The diagnostic abilities of these methods to discriminate <2 mm and ≥2 mm femoral stem subsidence were tested.The accuracy of the distal radius BMD and cortical-bone thickness of the distal radius were moderate (area under the curve, 0.737 and 0.726, respectively) in discriminating between <2 mm and ≥2 mm stem subsidence. Women with low cortical-bone thickness of the radius were more likely (odds ratio = 6.7; P = .002) to develop stem subsidence ≥2 mm. These subjects had lower total hip BMD (P = .007) and reduced thickness of the medial cortex of the proximal femur (P = .048) with lower middle (P < .001) and distal (P = .004) stem-to-canal fill ratios.Femoral stem stability and resistance to subsidence are sensitive to adequate bone stock and unaltered anatomy. DXA and pulse-echo ultrasonometry of the distal radius may help discriminate postmenopausal women at high risk of stem subsidence.
View details for DOI 10.1016/j.arth.2020.02.062
View details for Web of Science ID 000540694800020
View details for PubMedID 32205004
- RSA of the Symax hip stem ACTA ORTHOPAEDICA 2020; 91 (4): 497–99
Automated segmentation of acute stroke lesions using a data-driven anomaly detection on diffusion weighted MRI
JOURNAL OF NEUROSCIENCE METHODS
2020; 333: 108575
Successful delineation of lesions in acute ischemic strokes (AIS) is crucial for increasing the likelihood of good clinical outcome for the patient.We developed a fully automated method to localize and segment AIS lesions in variable locations for 192 multimodal 3D-magnetic resonance images (MRI) including 106 stroke and 86 healthy cases. The method works based on the Crawford-Howell t-test and comparison of stroke images to healthy controls. We then developed a classifier to discriminate the images into stroke or non-stroke categories following the lesion segmentation.The mean Dice similarity coefficient (DSC) for the test set was 0.50 ± 0.21 (min-max: 0.07-0.83) and mean net overlap was 0.66 ± 0.18 (min-max: 0.22-1). The experimental results for the classification of strokes from non-strokes showed mean accuracy, precision, sensitivity, and specificity of 73 %, 0.77 %, 84 %, and 69 %, respectively.The performance of our methods is comparable with previously published approaches based on machine learning and/or deep learning lesion segmentation techniques. However, most of the previously published methods have yielded low sensitivity, are computationally heavy, and difficult to interpret. The present approach is a significant improvement because it does not require high computation power and memory and can be implemented on a desktop workstation and integrated into the routine clinical diagnostic pipeline.The current method is straightforward, fast, and shows good agreement with the lesions identified by human experts.
View details for DOI 10.1016/j.jneumeth.2019.108575
View details for Web of Science ID 000515428500010
View details for PubMedID 31904391
Effect of Denosumab on Femoral Periprosthetic BMD and Early Femoral Stem Subsidence in Postmenopausal Women Undergoing Cementless Total Hip Arthroplasty.
2019; 3 (10): e10217
Antiresorptive denosumab is known to improve the quality and strength of cortical bone in the proximal femurs of osteoporotic women, but its efficacy in preventing periprosthetic bone loss and reducing femoral stem migration has not been studied in women undergoing cementless total hip arthroplasty. We conducted a single-center, randomized, double-blinded, placebo-controlled trial of 65 postmenopausal women with primary hip osteoarthritis and Dorr type A or B proximal femur anatomy. The patients randomly received subcutaneous injections of denosumab 60 mg or placebo once every 6 months for 12 months, starting 1 month before surgery. The primary endpoint was the change in bone mineral density (BMD) of the proximal femur (Gruen zone 7) at week 48, and the secondary endpoint was stem subsidence measured by radiostereometric analysis (RSA) at week 48. Exploratory endpoints included changes in BMDs of the contralateral hip, lumbar spine and distal radius, serum levels of bone turnover markers, walking speed, walking activity, patient-reported outcome measures, and radiographic assessment of stem osseointegration. The participants underwent vertebral-fracture assessment in an extension safety study at 3 years. Denosumab significantly decreased bone loss in the medial femoral neck (zone 7) and increased periprosthetic BMD in the greater trochanteric region (zone 1) and lesser trochanteric region (zone 6). Denosumab did not reduce temporary femoral stem migration. The migration occurred mainly during the settling period (0 to 12 weeks) after implantation of the prosthesis. All of the stems osseointegrated, as evaluated by RSA and radiographs. There were no intergroup differences in functional recovery. Discontinuation of denosumab did not lead to any adverse events. In conclusion, denosumab increased periprosthetic BMD in the clinically relevant regions of the proximal femur, but the treatment response was not associated with any reduction of initial stem migration. © 2019 The Authors. JBMR Plus published by Wiley Periodicals, Inc. on behalf of the American Society for Bone and Mineral Research.
View details for DOI 10.1002/jbm4.10217
View details for PubMedID 31687650
View details for PubMedCentralID PMC6820573
Simplified Automated Segmentation of Acute Ischemic Stroke Lesions from Multimodal MRI: A knowledge-based learning approach
View details for Web of Science ID 000569982800317
Is Model-based Radiostereometric Analysis Suitable for Clinical Trials of a Cementless Tapered Wedge Femoral Stem?
CLINICAL ORTHOPAEDICS AND RELATED RESEARCH
2016; 474 (10): 2246–53
In clinical trials of THA, model-based radiostereometric analysis (RSA) techniques may be less precise than conventional marker-based RSA for measurement of femoral stem rotation. We verified the accuracy and clinical precision of RSA based on computer-aided design models of a cementless tapered wedge femoral stem.We asked: (1) Is the accuracy of model-based RSA comparable to that of marker-based RSA? (2) What is the clinical precision of model-based RSA?Model-based RSA was performed using combined three-dimensional computer-aided design models of the stem and head provided by the implant manufacturer. The accuracy of model-based RSA was compared with that of marker-based RSA in a phantom model using micromanipulators for controlled translation in three axes (x, y, z) and rotation around the y axis. The clinical precision of model-based RSA was evaluated by double examinations of patients who had arthroplasties (n = 24) in an ongoing trial. The clinical precision was defined as being at an acceptable level if the number of patients needed for a randomized trial would not differ from a trial done with conventional marker-based RSA (15-25 patients per group).The accuracy of model-based RSA was 0.03 mm for subsidence (translation along the y axis) (95% CI for the difference between RSA measurements and actual displacement measured with micrometers, -0.03-0.00) and 0.39° for rotation around the y axis (95% CI, -0.41 to -0.06). The accuracy of marker-based RSA was 0.06 mm for subsidence (95% CI, -0.04-0.01; p = 0.728 compared with model-based RSA) and 0.18° for the y axis rotation (95% CI, -0.23 to -0.07; p = 0.358). The clinical precision of model-based RSA was 0.14 mm for subsidence (95% CI for the difference between double examinations, -0.02-0.04) and 0.79° for the y axis rotation (95% CI, -0.16-0.18).The accuracy of model-based RSA for measurement of the y axis rotation was not quite as high as that of marker-based RSA, but its clinical precision is at an acceptable level.Model-based RSA may be suitable for clinical trials of cementless tapered wedge femoral stem designs.
View details for DOI 10.1007/s11999-016-4930-0
View details for Web of Science ID 000382930700031
View details for PubMedID 27334320
View details for PubMedCentralID PMC5014820
AvanTomography: A Compact Module for Positron Emission Mammography
IEEE. 2015: 52–57
View details for Web of Science ID 000380610600010