Alessan­dro Crimi

Brain Graph net­works between bio­mark­ers, graph con­vo­lu­tion­al net­works, and causality

Abstract

Brain con­nec­tiv­i­ty refers to the approach of rep­re­sent­ing dif­fer­ent aspects of the brain (struc­tur­al con­nec­tions, cor­re­la­tion of blood con­cen­tra­tion, gene expres­sions, etc). As a bio­mark­er, brain con­nec­tiv­i­ty pro­vides valu­able insights into the brain. By map­ping and ana­lyz­ing the intri­cate net­work of con­nec­tions between dif­fer­ent brain regions, researchers can iden­ti­fy aber­rant con­nec­tiv­i­ty pat­terns asso­ci­at­ed with var­i­ous neu­ro­log­i­cal dis­or­ders such as Alzheimer’s dis­ease, schiz­o­phre­nia, and autism spec­trum dis­or­ders. These find­ings aid in ear­ly detec­tion, diag­no­sis, and mon­i­tor­ing of these con­di­tions, allow­ing for tar­get­ed inter­ven­tions and per­son­al­ized treat­ment strategies.

Fur­ther­more, the advent of graph con­vo­lu­tion­al net­works has rev­o­lu­tion­ized the field of brain con­nec­tiv­i­ty analy­sis. GCNs uti­lize graph the­o­ry and deep learn­ing tech­niques to extract mean­ing­ful fea­tures from brain con­nec­tiv­i­ty data. By lever­ag­ing the rich infor­ma­tion encod­ed in brain net­works, GCNs enable accu­rate clas­si­fi­ca­tion, pre­dic­tion, and under­stand­ing of brain-relat­ed phe­nom­e­na. They have been suc­cess­ful­ly applied in tasks such as brain image seg­men­ta­tion, func­tion­al con­nec­tiv­i­ty analy­sis, and brain net­work com­par­i­son, pro­vid­ing nov­el insights into the com­plex dynam­ics of the human brain.

More­over. the com­plex and non­lin­ear nature of brain dynam­ics often makes it dif­fi­cult to estab­lish direct causal links. Researchers must care­ful­ly con­sid­er con­found­ing fac­tors, such as net­work topol­o­gy and feed­back loops, to ensure the valid­i­ty of their mod­els. Advanced sta­tis­ti­cal and com­pu­ta­tion­al meth­ods, such as Bayesian net­works and struc­tur­al equa­tion mod­el­ing, are employed to address these challenges. 

Biog­ra­phy

Dr. Alessan­dro Cri­mi is a bio­med­ical engi­neer and health econ­o­mist who alter­nat­ed his career between neu­roimag­ing and health­care man­age­ment in low-income countries.

He was born in Italy. After com­plet­ing his stud­ies in engi­neer­ing at the Uni­ver­si­ty of Paler­mo, he obtained a PhD in machine learn­ing applied for med­ical imag­ing by the Uni­ver­si­ty of Copen­hagen, and an MBA in health­care man­age­ment by the Uni­ver­si­ty of Basel. 

Alessan­dro worked as post-doc­tor­al researcher at the French Insti­tute for Research in Com­put­er Sci­ence (INRIA), Tech­ni­cal School of Switzer­land (ETH-Zurich), Ital­ian Insti­tute for Tech­nol­o­gy (IIT), and Uni­ver­si­ty Hos­pi­tal of Zurich. In those insti­tutes he made sig­nif­i­cant con­tri­bu­tions in the field of com­pu­ta­tion­al neuroscience. 

The post-doc­tor­al years at Euro­pean insti­tutes were alter­nat­ed by peri­od lived in Ghana and oth­er Sub-Saha­ran coun­tries where Dr. Cri­mi taught and car­ried out in-field projects about health­care man­age­ment. He taught for 8 years at the African Insti­tute for Math­e­mat­i­cal Sci­ence (AIMS) in Ghana and South Africa the course of machine learn­ing in med­i­cine, where he also super­vised numer­ous MSc theses.

The projects he con­duct­ed in Sub-Saha­ran Africa were relat­ed to pre­na­tal care, dia­betes, malar­ia and HIV man­age­ment using nov­el tech­nolo­gies as machine learn­ing, image pro­cess­ing and social engineering.

Since 2021, Dr. Cri­mi moved to the Cen­ter for com­pu­ta­tion­al med­i­cine in Poland, where is a prin­ci­pal inves­ti­ga­tor, he had to lim­it his engage­ments with AIMS to MSc the­sis supervision.

Dr. Cri­mi is cur­rent­ly the head of the neu­roimag­ing lab at Sano work­ing on find­ing nov­el bio­mark­ers relat­ed to brain dis­eases with new tech­nolo­gies as machine learn­ing and quan­tum computing.

More­over, he is cur­rent­ly involved in activ­i­ties for tech­nol­o­gy trans­fer to train immi­grant and women with chil­dren towards entrepreneurship.