Marieatou Daniels

Coding can be very frus­tra­ting, but when you’ve spent days working on a problem, the feeling you get when you find the so­lu­ti­on is in­de­scri­bable. I love solving pro­blems. I dis­co­ve­r­ed my passion for complex codes during an in­ternship at the Max Planck In­sti­tu­te for Human Co­gni­ti­ve and Brain Sci­en­ces in Leipzig. I am con­vin­ced the future of neu­ro­sci­ence lies in com­pu­ter-aided methods, and ad­van­ces in machine lear­ning are opening up bound­less pos­si­bi­li­ties.

There is much talk no­wa­days about per­so­na­li­zed me­di­ci­ne. In this area, al­go­rithms are helping us create precise patient-spe­ci­fic tre­at­ment models. They con­den­se the com­ple­xi­ty of the brain with its bil­li­ons of neurons and tril­li­ons of syn­ap­ses.

Right now, I’m excited about the pos­si­bi­li­ty of pro­gramming a self-su­per­vi­sed neural network that can de­scri­be be­ha­vi­or pat­terns in la­bo­ra­to­ry mice au­to­no­mously. That’s what I’m working on at the moment. Imagine someone is con­duc­ting a study on the impacts of a par­ti­cu­lar gene in epi­le­psy in order to find new tre­at­ment ap­proa­ches. A mouse that has a genetic mu­ta­ti­on as­so­cia­ted with epi­le­psy is likely to exhibit be­ha­vio­ral changes in an animal ex­pe­ri­ment.

The neural network I worked on with others at Harvard Medical School under Caleb Weinreb is capable of re­co­gni­zing and clas­si­fy­ing such changes based on video re­cord­ings. Until now, you would have needed experts to go through each video, frame by frame, and high­light be­ha­vi­or pat­terns. This time-con­suming work is now done by our al­go­rithms. In the future, they will even iden­ti­fy new be­ha­vi­or pat­terns because they can spot tiny re­cur­ring mo­ve­ments.

Through my work I hope to con­tri­bu­te to the de­ve­lop­ment of the­ra­peu­tic ap­proa­ches for neu­ro­de­ge­ne­ra­ti­ve di­sea­ses. A very close re­la­ti­ve of mine suffers from mul­ti­ple scle­ro­sis, and I know how de­vas­ta­ting it is. I hope to be able to give pa­ti­ents and their loved ones new options. Code can help close the gap between bench and bedside and speed up the trans­la­ti­on of sci­en­ti­fic dis­co­ve­ries into tre­at­ments for pa­ti­ents.

In the future, I would like to in­ves­ti­ga­te the link between ADHD and crea­ti­vi­ty and use graph theory to un­der­stand how dif­fe­rent regions of the brain are con­nec­ted and how this is dif­fe­rent in ADHD. I am also in­te­rested in ana­ly­zing neural data, for example neuron spike pat­terns in people with neu­ro­de­ge­ne­ra­ti­ve di­sea­ses. And I’m in­te­rested in brain-machine in­ter­faces too.

My head is full of ideas. I imagine my brain as a big mind map where each topic is linked to around 20 others. This lets me ap­proach pro­blems from unusual angles and find crea­ti­ve so­lu­ti­ons. In my free time, I like to draw and I love making col­la­ges. Col­la­ges don’t have strict bounda­ries, rules, or cons­traints. You can simply have ideas that trigger more ideas.