Premium
Intra‐ and inter‐rater reliability of brain's activation area measurement on imagej: An easy and low‐cost method for cognitive and neurophysiological studies
Author(s) -
MonteiroJunior Renato Sobral,
Santos Iaggo Raphael David,
Machado Frederico Sander
Publication year - 2021
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.058591
Subject(s) - pixel , computer science , region of interest , artificial intelligence , cartography , computer vision , geography
Background Neuroimage is an important resource in cognitive neuroscience 1 but its analysis is complex and requires reliable and affordable procedures. The present study sought to determine the intra‐ and inter‐rater reliability of brain´s activation area of neuroimages obtained derived from an integrated EEG system equipment. Method The study follows the Declaration of Helsinki. Fifteen cerebral cortex images were obtained through Neurometry v6 (Bioevolution ® , Brazilian National Health Surveillance Agency registration 81403519002) 2,3 from 15 individuals above 60 years, who consented to participate of the study. This device consists of a 19‐channel EEG (10‐20 system). Amplitude of brain waves (delta, theta, alpha, beta, and gamma) are automatically processed, filtered and analyzed with Math Library MtxVec, LAPACK and Math Kernel BLAS with codes Streaming SIMD (SSE2, SSE3 and SSE4) 2 . The neuroimages display different colors according to the amplitude of brain waves: brown/blue (low activation), green/yellow (moderate activation), and orange/red (high activation) 2 . Images retrieved were exported and analyzed with the ImageJ software (National Institutes of Health, Bethesda, Maryland, USA) in 3 phases: 1 st An experienced researcher (RSMJ) standardized Broadmann´s area as the region of interest (ROI); 2 nd The researcher specified the brain´s activation area in accordance with colored regions (brown, blue, green, yellow, orange, and red) in the Broadmann´s area 10 (BA10) covering 1,391 pixels. Colors were covered using 2 ImageJ tools (polygon and freehand selections). In order to calculate the colored area of interest, pixels quantification was performed with a 400% magnification filter (Figure 1). The procedure was repeated 24h later for the intra‐rater reliability; 3 rd An inexperienced assessor (IRDS) independently performed the same procedure of the 2 nd phase. Intraclass Correlation Coefficients (ICC 3,1 and ICC 3,k ) 4 were calculated to determine the intra‐ and inter‐rater reproducibility. Typical Error of Measurement (TEM= SD ∑ difference between measures÷√2 ) 5 was also calculated to quantify typical deviations from the right measure. Results Intra‐ and inter‐rater reliability were excellent (ICC=0.99, p<0.01) and good (ICC=0.82, p<0.01), respectively (Table 1). TEM varied from 4 to 17%. Conclusion Quantification of the brain´s activation area of neuroimages derived from Neurometry v6 using the ImageJ ROI is highly reliable, independently of the assessor´s experience.