Giulia Baldazzi

Dr. Giulia Baldazzi

University of Cagliari, Cagliari, Sardegna, Italy
+1
orcid

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 December 2020 in Journal of Neural Engineering
Verified
Journal Article
Open access
Giulia BaldazziGiulia Baldazzi
Giulia Baldazzi
Giuliana SolinasGiuliana Solinas
Giuliana Solinas
JMJaume del Valle Macià
Jaume del Valle Macià
MBMassimo Barbaro
Massimo Barbaro
SMSilvestro Micera
Silvestro Micera
Danilo PaniDanilo Pani
Danilo Pani
Objective. Among the different approaches for denoising neural signals, wavelet-based methods are widely used due to their ability to reduce in-band noise. All wavelet denoising algorithms have a common structure, but their effectiveness strongly depends on several implementation choices, including the mother wavelet, the decomposition level, the threshold definition, and the way it is applied (i.e. the thresholding). In this work, we investigated these factors to quantitatively assess their effects on neural signals in terms of noise reduction and morphology preservation, which are important when spike sorting is required downstream. Approach. Based on the spectral characteristics of the neural signal, according to the sampling rate of the signals, we considered two possible decomposition levels and identified the best-performing mother wavelet. Then, we compared different threshold estimation and thresholding methods and, for the best ones, we also evaluated their effect on clearing the approximation coefficients. The assessments were performed on synthetic signals that had been corrupted by different types of noise and on a murine peripheral nervous system dataset, both of which were sampled at about 16 kHz. The results were statistically analysed in terms of their Pearson’s correlation coefficients, root-mean-square errors, and signal-to-noise ratios. Main results. As expected, the wavelet implementation choices greatly influenced the processing performance. Overall, the Haar wavelet with a five-level decomposition, hard thresholding method, and the threshold proposed by Han et al (2007) achieved the best outcomes. Based on the adopted performance metrics, wavelet denoising with these parametrizations outperformed conventional 300–3000 Hz linear bandpass filtering. Significance. These results can be used to guide the reasoned and accurate selection of wavelet denoising implementation choices in the context of neural signal processing, particularly when spike-morphology preservation is required.

1 October 2020 in Computer Methods and Programs in Biomedicine
Verified
Journal Article
Open access
Giulia BaldazziGiulia Baldazzi
Giulia Baldazzi
ESEleonora Sulas
Eleonora Sulas
MUMonica Urru
Monica Urru
RTRoberto Tumbarello
Roberto Tumbarello
LRLuigi Raffo
Luigi Raffo
Danilo PaniDanilo Pani
Danilo Pani
The detection of a clean and undistorted foetal electrocardiogram (fECG) from non-invasive abdominal recordings is an open research issue. Several physiological and instrumental noise sources hamper this process, even after that powerful fECG extraction algorithms have been used. Wavelet denoising is widely used for the improvement of the SNR in biomedical signal processing. This work aims to systematically assess conventional and unconventional wavelet denoising approaches for the post-processing of fECG signals by providing evidence of their effectiveness in improving fECG SNR while preserving the morphology of the signal of interest. The stationary wavelet transform (SWT) and the stationary wavelet packet transform (SWPT) were considered, due to their different granularity in the sub-band decomposition of the signal. Three thresholds from the literature, either conventional (Minimax and Universal) and unconventional, were selected. To this aim, the unconventional one was adapted for the first time to SWPT by trying different approaches. The decomposition depth was studied in relation to the characteristics of the fECG signal. Synthetic and real datasets, publicly available for benchmarking and research, were used for quantitative analysis in terms of noise reduction, foetal QRS detection performance and preservation of fECG morphology. The adoption of wavelet denoising approaches generally improved the SNR. Interestingly, the SWT methods outperformed the SWPT ones in morphology preservation (p<0.04) and SNR (p<0.0003), despite their coarser granularity in the sub-band analysis. Remarkably, the Han et al. threshold, adopted for the first time for fECG processing, provided the best quality improvement (p<0.003). The findings of our systematic analysis suggest that particular care must be taken when selecting and using wavelet denoising for non-invasive fECG signal post-processing. In particular, despite the general noise reduction capability, signal morphology can be significantly altered on the basis of the parameterization of the wavelet methods. Remarkably, the adoption of a finer sub-band decomposition provided by the wavelet packet was not able to improve the quality of the processing.

23 March 2022 in Frontiers in Bioengineering and Biotechnology
Verified
Journal Article
Open access
Spanu AndreaSpanu Andrea
Spanu Andrea
AMAntonello Mascia
Antonello Mascia
Giulia BaldazziGiulia Baldazzi
Giulia Baldazzi
BFBenji Fenech-Salerno
Benji Fenech-Salerno
FTFelice Torrisi
Felice Torrisi
DPDanilo Pani
Danilo Pani
A breathable tattoo electrode for bio-potential recording based on a Parylene C nanofilm is presented in this study. The proposed approach allows for the fabrication of micro-perforated epidermal submicrometer-thick electrodes that conjugate the unobtrusiveness of Parylene C nanofilms and the very important feature of breathability. The electrodes were fully validated for electrocardiography (ECG) measurements showing performance comparable to that of conventional disposable gelled Ag/AgCl electrodes, with no visible negative effect on the skin even many hours after their application. This result introduces interesting perspectives in the field of epidermal electronics, particularly in applications where critical on-body measurements are involved.

27 April 2023 in Scientific Reports
Verified
Journal Article
Open access
PDF file
Giulia BaldazziGiulia Baldazzi
Giulia Baldazzi
MOMarco Orrù
Marco Orrù
GVGraziana Viola
Graziana Viola
DPDanilo Pani
Danilo Pani
Nowadays, catheter-based ablation in patients with post-ischemic ventricular tachycardia (VT) is performed in arrhythmogenic sites identified by electrophysiologists by visual inspection during electroanatomic mapping. This work aims to present the development of machine learning tools aiming at supporting clinicians in the identification of arrhythmogenic sites by exploiting innovative features that belong to different domains. This study included 1584 bipolar electrograms from nine patients affected by post-ischemic VT. Different features were extracted in the time, time scale, frequency, and spatial domains and used to train different supervised classifiers. Classification results showed high performance, revealing robustness across the different classifiers in terms of accuracy, true positive, and false positive rates. The combination of multi-domain features with the ensemble tree is the most effective solution, exhibiting accuracies above 93% in the 10-time 10-fold cross-validation and 84% in the leave-one-subject-out validation. Results confirmed the effectiveness of the proposed features and their potential use in a computer-aided system for the detection of arrhythmogenic sites. This work demonstrates for the first time the usefulness of supervised machine learning for the detection of arrhythmogenic sites in post-ischemic VT patients, thus enabling the development of computer-aided systems to reduce operator dependence and errors, thereby possibly improving clinical outcomes.

1 July 2020
Verified
Conference Paper
Closed access
AZA. Zedda
A. Zedda
EGE. Gusai
E. Gusai
Marco CarusoMarco Caruso
Marco Caruso
SBS. Bertuletti
S. Bertuletti
Giulia BaldazziGiulia Baldazzi
Giulia Baldazzi
DPD. Pani
D. Pani
After a cerebral stroke, survivors need to follow a neurorehabilitation program including exercises to be executed under a therapist’s supervision or autonomously. Technological solutions are needed to support the early discharge of the patients just after the primary hospital treatments, by still providing an adequate level of rehabilitation. The DoMoMEA Project proposes a fully-wearable m-health solution able to administer a neurorehabilitation therapy in the patient's home or every other place established by the patient for a rehabilitation session. The exploitation of magneto-inertial measurement units only, wirelessly connected to an Android-operated device, provides robustness to different operating conditions and immunity to optical occlusion problems, compared to RGB-D cameras. Patients’ engagement is fostered by the exploitation of the exergame version of the ten rehabilitation exercises, implemented in Unity 3D. Store-and-forward telemonitoring features, supported by cloud-based storage and by a web application accessible from anywhere by medical personnel and patients, enable constant transparent monitoring of the rehabilitation progresses. The clinical trial of the DoMoMEA telerehabilitation system will involve 40 post-stroke patients with mild impairment and will start as soon as the restrictions due to the COVID-19 pandemic will allow to enroll patients.

13 September 2024 in SLEEP
Verified
Journal Article
Open access
PDF file
VGValentina Gnoni
Valentina Gnoni
LTLudovica Tamburrino
Ludovica Tamburrino
Giulia BaldazziGiulia Baldazzi
Giulia Baldazzi
DUDaniele Urso
Daniele Urso
SZStefano Zoccolella
Stefano Zoccolella
GLGiancarlo Logroscino
Giancarlo Logroscino
Study Objectives: Sleep disorders have been recognized as an integral component of the clinical syndrome in several neurodegenerative diseases, including Alzheimer’s disease (AD). However, limited data exist for rarer types of neurodegenerative diseases, such as behavioral variant frontotemporal dementia (bvFTD). This study aims to analyze EEG power spectra and sleep stage transitions in bvFTD patients, hypothesizing that bvFTD may show distinctive sleep stage transitions compared to patients with AD. Methods: Eighteen probable bvFTD patients and 18 age- and sex-matched probable patients with AD underwent overnight polysomnography (PSG) and completed sleep disorders questionnaires. Sleep questionnaires, full-night EEG spectra, and sleep stage transition indexes were compared between groups. Results: bvFTD patients had higher Insomnia Severity Index (ISI) scores (95% confidence intervals [CI]: 0, 5) and reported poorer sleep quality than AD patients (p < .01). Compared to AD, bvFTD patients showed higher N1 percentage (95% CI: 0.1, 6), lower N3 percentage (95% CI: −13.6, −0.6), higher sleep–wake transitions (95% CI: 1.49, 8.86) and N1 sleep–wake transitions (95% CI: 0.32, 6.1). EEG spectral analysis revealed higher spectral power in bvFTD compared to patients with AD in faster rhythms, especially sigma rhythm, across all sleep stages. In bvFTD patients, sleep–wake transitions were positively associated with ISI. Conclusions: Patients with bvFTD present higher rates of transitions between wake and sleep than patients with AD. The increased frequency of sleep transitions indicates a higher degree of sleep instability in bvFTD, which may reflect an imbalance in sleep–wake-promoting systems. Sleep stage transitions analysis may provide novel insights into the sleep alterations of patients with bvFTD.