Added Value of QEEG for the Differential Diagnosis of Common Forms of Dementia

Quantitative EEG in the Differential Diagnosis of Dementia Subtypes

OBJECTIVE

Most neurodegenerative dementias present with substantial overlap in clinical features. Therefore, differential diagnosis is often a challenging task necessitating costly and sometimes invasive diagnostic procedures. A promising, non-invasive and cost-effective method is the widely available electroencephalography (EEG).

METHODS

Twenty-three subjects with Alzheimer's disease (AD), 28 subjects with dementia with Lewy bodies (DLB), 15 subjects with frontotemporal dementias (FTDs), and 22 healthy controls (HC) were enrolled. Nineteen channel computerized EEG recordings were acquired. Mean relative powers were calculated using the standard frequency bands. Theta/alpha ratio (TAR), theta/beta ratio (TBR), a spectral index of (alpha + beta)/(theta + delta) and an alpha reactivity index (alpha in eyes-open condition/alpha in eyes-closed condition) were also calculated. Receiver operating characteristic (ROC) analyses were performed to assess diagnostic accuracy.

RESULTS

For the comparison of EEG measures across groups, we performed a multivariate ANOVA followed by univariate ANOVAs controlling for the effects of age, with post hoc tests. Theta power and TBR were increased in DLB compared to other groups. Alpha power was decreased in DLB compared to HC and FTD; and in AD compared to FTD. Beta power was decreased in DLB compared to AD and HC. Furthermore, regional analyses demonstrated a unique pattern of theta power increase in DLB; affecting frontal, central, parietal, occipital, and temporal regions. In AD, theta power increased compared to HC in parietal, occipital, and right temporal regions. TAR was increased in DLB compared to other groups; and in AD compared to HC. Finally, alpha reactivity index was higher in DLB compared to HC and FTD. In AD, EEG slowing was associated with cognitive impairment, while in DLB, this was associated with higher DLB characteristics. In the ROC analyses to distinguish DLB from FTD and AD, measures of EEG slowing yielded high area under curve values, with good specificities. Also, decreased alpha reactivity could distinguish DLB from FTD with good specificity. EEG slowing in DLB showed a diffuse pattern compared to AD, where a posterior and temporal slowing predominated.

CONCLUSION

We showed that EEG slowing was satisfactory in distinguishing DLB patients from AD and FTD patients. Notably, this slowing was a characteristic finding in DLB patients, even at early stages, while it paralleled disease progression in AD. Furthermore, EEG slowing in DLB showed a diffuse pattern compared to AD, where a posterior and temporal slowing predominated. These findings align with the previous evidence of the diencephalic dysfunction in DLB.

Machine and deep learning algorithms for classifying different types of dementia: A literature review

The cognitive impairment known as dementia affects millions of individuals throughout the globe. The use of machine learning (ML) and deep learning (DL) algorithms has shown great promise as a means of early identification and treatment of dementia. Dementias such as Alzheimer's Dementia, frontotemporal dementia, Lewy body dementia, and vascular dementia are all discussed in this article, along with a literature review on using ML algorithms in their diagnosis. Different ML algorithms, such as support vector machines, artificial neural networks, decision trees, and random forests, are compared and contrasted, along with their benefits and drawbacks. As discussed in this article, accurate ML models may be achieved by carefully considering feature selection and data preparation. We also discuss how ML algorithms can predict disease progression and patient responses to therapy. However, overreliance on ML and DL technologies should be avoided without further proof. It's important to note that these technologies are meant to assist in diagnosis but should not be used as the sole criteria for a final diagnosis. The research implies that ML algorithms may help increase the precision with which dementia is diagnosed, especially in its early stages. The efficacy of ML and DL algorithms in clinical contexts must be verified, and ethical issues around the use of personal data must be addressed, but this requires more study.

Quantitative electroencephalography in cerebral amyloid angiopathy

OBJECTIVE

We investigated whether quantitative electroencephalography (qEEG) correlates with cognition and cortical superficial siderosis (cSS) in cerebral amyloid angiopathy.

METHODS

We included patients with sporadic (sCAA) and hereditary Dutch-type CAA (D-CAA). Spectral measures and the phase lag index (PLI) were analyzed on qEEG. Cognition was assessed with the MoCA and cSS presence was scored on 3T-MRI. Linear regression analyses were performed to investigate these qEEG measures and cognition. Independent samples T-tests were used to analyze the qEEG measure differences between participants with and without cSS.

RESULTS

We included 92 participants (44 D-CAA; 48 sCAA). A lower average peak frequency (β[95 %CI] = 0.986[0.252-1.721]; P = 0.009) and a higher spectral ratio (β[95 %CI] = -0.918[-1.761--0.075]; P = 0.033) on qEEG correlated with a lower MoCA score, irrespective of a history of symptomatic intracerebral hemorrhage (sICH). The PLI showed no correlation to the MoCA. qEEG slowing was not different in those with or without cSS.

CONCLUSIONS

Spectral qEEG (but not PLI) reflects cognitive performance in patients with CAA with and without a history of sICH. We found no association between qEEG slowing and cSS.

SIGNIFICANCE

qEEG could be a valuable biomarker, especially in challenging cognitive testing situations in CAA, and a potential predictive tool in future studies.

Radiographers' perceptions of the experiences of patients with dementia attending the radiology department

BACKGROUND

Patients with dementia frequently present to the Radiology Department. However, stigmas have been recorded amongst radiographers surrounding imaging persons with dementia (PwD). This study aimed to investigate the impact of PwD attending the Radiology Department for imaging and the resultant effects to all patients, radiographers, and the Department from the perspectives of the examining radiographer.

METHODS

A paper-based questionnaire of radiographers' perceptions and experiences of individual examinations 'termed an interaction form' was created and made available in a public hospital in Ireland for a period of eight weeks. Radiographers completed the interaction form collecting data regarding individual imaging examinations of PwD. The form comprised sixteen closed and one open-ended question on the radiographers' individual perspectives of PwDs' abilities and distress levels, carers and comforters and their role in the examination, what the radiographer found helpful in the interaction, and any adverse events. Data were analysed using a combination of descriptive analysis and thematic content analysis.

RESULTS

Thirty-three interaction forms were completed by the participating radiographers. The modality most commonly represented in the survey was general X-ray (58%). Radiographers reported 84% of examinations for PwD required extra time, with 27% of examinations required repeat imaging and 69% of patients appeared distressed. A carer helped facilitate the completion of 77% of examinations. Qualitative data indicated that distractive and communicative techniques were used by radiographers to make the patient feel more comfortable and help with examination success.

CONCLUSION

PwD often require more time for radiological examinations, they often need repeat imaging and re-scheduling of an examination at a more appropriate time. These factors need to be considered when scheduling and performing radiological examinations. Patient distress was frequently encountered, this area may benefit from further research and dedicated practitioner training which could help drive improvements in patient experience.

EEG Correlates in the 3 Variants of Primary Progressive Aphasia

BACKGROUND AND OBJECTIVES

The 3 clinical presentations of primary progressive aphasia (PPA) reflect heterogenous neuropathology, which is difficult to be recognized in vivo. Resting-state (RS) EEG is promising for the investigation of brain electrical substrates in neurodegenerative conditions. In this study, we aim to explore EEG cortical sources in the characterization of the 3 variants of PPA.

METHODS

This is a cross-sectional, single-center, memory center-based cohort study. Patients with PPA and healthy controls were consecutively recruited at the Neurology Unit, IRCCS San Raffaele Scientific Institute (Milan, Italy). Each participant underwent an RS 19-channel EEG. Using standardized low-resolution brain electromagnetic tomography, EEG current source densities were estimated at voxel level and compared among study groups. Using an RS functional MRI-driven model of source reconstruction, linear lagged connectivity (LLC) values within language and extra-language brain networks were obtained and analyzed among groups.

RESULTS

Eighteen patients with logopenic PPA variant (lvPPA; mean age = 72.7 ± 6.6; % female = 52.4), 21 patients with nonfluent/agrammatic PPA variant (nfvPPA; mean age = 71.7 ± 8.1; % female = 66.6), and 9 patients with semantic PPA variant (svPPA; mean age = 65.0 ± 6.9; % female = 44.4) were enrolled in the study, together with 21 matched healthy controls (mean age = 69.2 ± 6.5; % female = 57.1). Patients with lvPPA showed a higher delta density than healthy controls (p < 0.01) and patients with nfvPPA (p < 0.05) and svPPA (p < 0.05). Patients with lvPPA also displayed a greater theta density over the left posterior hemisphere (p < 0.01) and lower alpha2 values (p < 0.05) over the left frontotemporal regions than controls. Patients with nfvPPA showed a diffuse greater theta density than controls (p < 0.05). LLC was altered in all patients relative to controls (p < 0.05); the alteration was greater at slow frequency bands and within language networks than extra-language networks. Patients with lvPPA also showed greater LLC values at theta band than patients with nfvPPA (p < 0.05).

DISCUSSION

EEG findings in patients with PPA suggest that lvPPA-related pathology is associated with a characteristic disruption of the cortical electrical activity, which might help in the differential diagnosis from svPPA and nfvPPA. EEG connectivity was disrupted in all PPA variants, with distinct findings in disease-specific PPA groups.

CLASSIFICATION OF EVIDENCE

This study provides Class IV evidence that EEG analysis can distinguish PPA due to probable Alzheimer disease from PPA due to probable FTD from normal aging.

Quantitative electroencephalography (qEEG), apolipoprotein A-I (APOA-I), and apolipoprotein epsilon 4 (APOE ɛ4) alleles for the diagnosis of mild cognitive impairment and Alzheimer's disease

INTRODUCTION

Alzheimer's disease (AD) is the most common type of dementia. Amnestic mild cognitive impairment (aMCI), a pre-dementia stage is an important stage for early diagnosis and intervention. This study aimed to investigate the diagnostic value of qEEG, APOA-I, and APOE ɛ4 allele in aMCI and AD patients and found the correlation between qEEG (Delta + Theta)/(Alpha + Beta) ratio (DTABR) and different cognitive domains.

METHODS

All participants were divided into three groups: normal controls (NCs), aMCI, and AD, and all received quantitative electroencephalography (qEEG), neuropsychological scale assessment, apolipoprotein epsilon 4 (APOE ɛ4) alleles, and various blood lipid indicators. Different statistical methods were used for different data.

RESULTS

The cognitive domains except executive ability were all negatively correlated with DTABR in different brain regions while executive ability was positively correlated with DTABR in several brain regions, although without statistical significance. The consequences confirmed that the DTABR of each brain area were related to MMSE, MoCA, instantaneous memory, and the language ability (p < 0.05), and the DTABR in the occipital area was relevant to all cognitive domains (p < 0.01) except executive function (p = 0.272). Also, occipital DTABR was most correlated with language domain when tested by VFT with a moderate level (r = 0.596, p < 0.001). There were significant differences in T3, T5, and P3 DTABR between both AD and NC and aMCI and NCs. As for aMCI diagnosis, the maximum AUC was achieved when using T3 combined with APOA-I and APOE ε4 (0.855) and the maximum AUC was achieved when using T5 combined with APOA-I and APOE ε4 (0.889) for AD diagnosis.

CONCLUSION

These findings highlight that APOA-I, APOE ɛ4, and qEEG play an important role in aMCI and AD diagnosis. During AD continuum, qEEG DTABR should be taken into consideration for the early detection of AD risk.

Normative brain mapping using scalp EEG and potential clinical application

A normative electrographic activity map could be a powerful resource to understand normal brain function and identify abnormal activity. Here, we present a normative brain map using scalp EEG in terms of relative band power. In this exploratory study we investigate its temporal stability, its similarity to other imaging modalities, and explore a potential clinical application. We constructed scalp EEG normative maps of brain dynamics from 17 healthy controls using source-localised resting-state scalp recordings. We then correlated these maps with those acquired from MEG and intracranial EEG to investigate their similarity. Lastly, we use the normative maps to lateralise abnormal regions in epilepsy. Spatial patterns of band powers were broadly consistent with previous literature and stable across recordings. Scalp EEG normative maps were most similar to other modalities in the alpha band, and relatively similar across most bands. Towards a clinical application in epilepsy, we found abnormal temporal regions ipsilateral to the epileptogenic hemisphere. Scalp EEG relative band power normative maps are spatially stable across time, in keeping with MEG and intracranial EEG results. Normative mapping is feasible and may be potentially clinically useful in epilepsy. Future studies with larger sample sizes and high-density EEG are now required for validation.

Source space connectomics of neurodegeneration: One-metric approach does not fit all

Brain functional connectivity in dementia has been assessed with dissimilar EEG connectivity metrics and estimation procedures, thereby increasing results' heterogeneity. In this scenario, joint analyses integrating information from different metrics may allow for a more comprehensive characterization of brain functional interactions in different dementia subtypes. To test this hypothesis, resting-state electroencephalogram (rsEEG) was recorded in individuals with Alzheimer's Disease (AD), behavioral variant frontotemporal dementia (bvFTD), and healthy controls (HCs). Whole-brain functional connectivity was estimated in the EEG source space using 101 different types of functional connectivity, capturing linear and nonlinear interactions in both time and frequency-domains. Multivariate machine learning and progressive feature elimination was run to discriminate AD from HCs, and bvFTD from HCs, based on joint analyses of i) EEG frequency bands, ii) complementary frequency-domain metrics (e.g., instantaneous, lagged, and total connectivity), and iii) time-domain metrics with different linearity assumption (e.g., Pearson correlation coefficient and mutual information). <10% of all possible connections were responsible for the differences between patients and controls, and atypical connectivity was never captured by >1/4 of all possible connectivity measures. Joint analyses revealed patterns of hypoconnectivity (patientsHCs) in both groups was mainly identified in frontotemporal regions. These atypicalities were differently captured by frequency- and time-domain connectivity metrics, in a bandwidth-specific fashion. The multi-metric representation of source space whole-brain functional connectivity evidenced the inadequacy of single-metric approaches, and resulted in a valid alternative for the selection problem in EEG connectivity. These joint analyses reveal patterns of brain functional interdependence that are overlooked with single metrics approaches, contributing to a more reliable and interpretable description of atypical functional connectivity in neurodegeneration.

Quantitative Electroencephalogram Standardization: A Sex- and Age-Differentiated Normative Database

We describe the utility of a standardized index (Z-score) in quantitative EEG (QEEG) capable of when referenced to a resting-state, sex- and age-differentiated QEEG normative database (ISB-NormDB). Our ISB-NormDB comprises data for 1,289 subjects (553 males, 736 females) ages 4.5 to 81 years that met strict normative data criteria. A de-noising process allowed stratification based on QEEG variability between normal healthy men and women at various age ranges. The ISB-NormDB data set that is stratified by sex provides a unique, highly accurate ISB-NormDB model (ISB-NormDB: ISB-NormDB-Male, ISB-NormDB-Female). To evaluate the trends and accuracy of the ISB-NormDB, we used actual data to compare Z-scores obtained through the ISB-NormDB with those obtained through a traditional QEEG normative database to confirm that basic trends are maintained in most bands and are sensitive to abnormal test data. Finally, we demonstrate the value of our standardized index of QEEG, and highlight it's capacity to minimize the confounding variables of sex and age in any analysis.

Application of Machine Learning to Electroencephalography for the Diagnosis of Primary Progressive Aphasia: A Pilot Study

Background. Primary progressive aphasia (PPA) is a neurodegenerative syndrome in which diagnosis is usually challenging. Biomarkers are needed for diagnosis and monitoring. In this study, we aimed to evaluate Electroencephalography (EEG) as a biomarker for the diagnosis of PPA. Methods. We conducted a cross-sectional study with 40 PPA patients categorized as non-fluent, semantic, and logopenic variants, and 20 controls. Resting-state EEG with 32 channels was acquired and preprocessed using several procedures (quantitative EEG, wavelet transformation, autoencoders, and graph theory analysis). Seven machine learning algorithms were evaluated (Decision Tree, Elastic Net, Support Vector Machines, Random Forest, K-Nearest Neighbors, Gaussian Naive Bayes, and Multinomial Naive Bayes). Results. Diagnostic capacity to distinguish between PPA and controls was high (accuracy 75%, F1-score 83% for kNN algorithm). The most important features in the classification were derived from network analysis based on graph theory. Conversely, discrimination between PPA variants was lower (Accuracy 58% and F1-score 60% for kNN). Conclusions. The application of ML to resting-state EEG may have a role in the diagnosis of PPA, especially in the differentiation from controls. Future studies with high-density EEG should explore the capacity to distinguish between PPA variants.