NBML_BioBank_AIChallange_20240430_01

Description:

The National Brain Mapping Laboratory, in line with its educational and research mission and as

a leading center in the development of Iranian Brain Mapping Biobank, is collaborating on the

Annual Iranian Artificial Intelligence Award challenge in the following areas:

1) Challenge Topic Design: To foster a competitive environment in the field of artificial

intelligence and promote the use of AI technologies to address pressing healthcare challenges,

the National Laboratory of Brain Mapping has selected the topic of "Intelligent Classification of

Normal and Abnormal Brain MRIs";

2) Challenge Data Provision: The data required for the challenge topic and for use in the annual

Iranian Artificial Intelligence Competition will be provided by the Iranian Brain Mapping

Biobank.

For this challenge, 50 de-identified brain structural MRI images have been provided in two

categories: normal (30 samples) and abnormal (20 samples). The abnormal samples represent

four disease categories: dementia, multiple sclerosis, neurovascular disease, and tumor.

The data have been anonymized and will be provided to participants with binary labels (normal

vs. abnormal).

 Image Characteristics:

 Brain Structural MRI acquired using a clinical protocol

 Recorded on a Siemens 1.5 Tesla MRI scanner

 Sequences included: T1-weighted spin-echo (T1_SE_TRA), T2-weighted turbo spin-

echo (T2_TSE_TRA), and T2-weighted inversion recovery dark-fluid

(T2_TIRM_TRA_DARK-FLUID)

 Axial Plane for all three sequences

 Binary labels (normal vs. abnormal) provided

 Efforts have been made to separate the data by observing homogeneity across these

parameters:

 Study Description: Brain MRI-C

 Series Description: t1_se_tra , t2_tse_tra , t2_tirm_tra_dark-fluid

 Slice_Thickness: 7.5

 Manufacturer : SIEMENS , Avanto

 Software_Versions: syngo MR B17

 Magnetic_Field_Strength: 1.5

 Flip Angle: 150/90

 Spacing_Between_Slices: 8.25

 Acquisition_Matrix: 0\256\155\0 , 0\384\129\0 , 0\256\133\0 , 0\256\167\0 ,

0\256\113\0 , 0\256\152\0 , 0\256\161\0 , 0\384\135\0