Page 220 - Handbook of Deep Learning in Biomedical Engineering Techniques and Applications

210 Chapter 7 Early detection and diagnosis using deep learning

Yet another way of detecting such autoimmune diseases is by
using the cluster approach of analysis. Cluster analysis is one of
the unsupervised methods of learning analysis in which data
are sent to a computer without having any sort of prior knowl-
edge, and the command analysis method is “whole bundle
similar things.” Without a specific conjecture, there is no target
variable with a hidden motive. The clustering techniques can be
represented in two methods. N constituent factors are nonhier-
archically classified into M clusters. The K-mean algorithm is a
typical technique of classifying data in K groups (clusters) that
reduce the variation of the difference between the distances of
each cluster. Four factors are categorized into four different vari-
ables in cluster analysis. Periodically grouped clusters multiple
group clusters in the consolidation process. Not only is the appli-
cation of clustering heterogeneous, but its techniques and idea-
tion can also be used to work with various different algorithms
(Fig. 7.5).

3.2 Alzheimer's disease
It is crucial to accurately diagnose AD early because the con-
sciousness of the severity and the progression risks allow patients
who are suffering from the condition to take prevention measures
before irreversible brain damages are shaped. This diagnosis
plays a significant role in patient care, especially at the early
stage. Present analysis of neuroimaging data, such as those
obtained methods such as magnetic resonance imaging (MRI),
positron emission tomography, functional MRI (fMRI), and diffu-
sion tensor imaging, requires handling by experts. AD is the most
common type of dementia that appears in persons over the age of
65 years. It is characterized by the progressive impairment
of cognitive and memory functions. To slow the progression of
dementia, timely treatment is vital. Such treatments rely on the
early diagnosis of AD and its prodromal stage, mild cognitive
impairment (MCI). Hence, what is required is a trustworthy and
reliable detection from brain imaging, and over that, a robust
diagnostic system, along with the aid of the analysis of neuroi-
maging data, will crave the path for a more instructive and infor-
mative approach. Diagnostic accuracy can also be potentially
increased through this. Before, the methods used to explore
neuroimaging biomarkers had a different structure, and their
base was built upon MUS (mass univariate statistics). It was inap-
propriately assumed by them that the actions of different regions
of the brain are not dependent upon each other.