for High Dimensional Prediction & Behavioral Phenotyping Isaac - - PowerPoint PPT Presentation

▶

Jan 29, 2023 234 likes •559 views

Machine Learning: A Framework for High Dimensional Prediction & Behavioral Phenotyping Isaac R. Galatzer-Levy, PhD Mindstrong Health NYU School of Medicine Disclosure Equity and salary from Mindstrong Health 1. Personalized medicine

SLIDE 1

Machine Learning: A Framework for High Dimensional Prediction & Behavioral Phenotyping

Isaac R. Galatzer-Levy, PhD Mindstrong Health NYU School of Medicine

SLIDE 2

Disclosure

Equity and salary from Mindstrong Health

SLIDE 3

1. Personalized medicine problem
Differential response to treatment
Variable & multidetermined treatment effect
Strong effects in subpopulations are washed out by averaging across responders and

non-responders

The ability to predict responders to Tx [x1…xn] can solve this problem

SLIDE 4

2. Phenotyping problem
Current clinical definitions
Heuristic
Heterogeneous
Distant from the behavioral and biological phenotypes used in basic and

translational models

Good treatments are lost in translation

SLIDE 5

Personalized medicine problem

1. 2. 3.

Breast cancer tissues Normal tissues

SLIDE 6

Basic principles of classification

Want to classify objects as boats and houses.

SLIDE 7

Basic principles of classification

All objects before the coast line are boats and all objects after the

coast line are houses.

Coast line serves as a decision surface that separates two classes.

SLIDE 8

Basic principles of classification

These boats will be misclassified as houses

SLIDE 9

Basic principles of classification

This house will be misclassified as a boat

SLIDE 10

High Dimensional Classification

SLIDE 11

SLIDE 12

SLIDE 13

SLIDE 14

SLIDE 15

SLIDE 16

Neural Networks for Deep Learning

SLIDE 17

Hot Dog Not Hot Dog

SLIDE 18

Phenotyping problem

Need for new ways to define clinical populations that move away

from heuristic description of mental health towards direct behavioral and biological phenotyping

Problem
High dimensional data
No clear clinical interpretation

SLIDE 19

Diagnostic Status

2 4 6 8 PTSD + PTSD - 2 4 6 8 PTSD + PTSD -

DSM-IV DSM 5 n= 79,794 6-17 symptoms n=636,120 6-20 symptoms

SLIDE 20

Digital Phenotyping

SLIDE 21

p n X= Orthogonal Basis

SLIDE 22

Now what?

SLIDE 23

SLIDE 24

Thank You

SLIDE 25

Latent Growth Mixture Modeling

T1 T2 T3 T4 S I Q C

SLIDE 26

Now what?

SLIDE 27

I,S,Q=>C

Resilience Chronic Recovery

Time

Latent Growth Mixture Modeling

SLIDE 28

Freezing

SLIDE 29

Freezing

SLIDE 30

Role of FKBP5 on Patterns of Fear Potentiated Startle During Extinction

Combined Iraq and Afghan War Vets and Grady ER Sample

n= 721

High FPS Extinguishers: RS1360780 TT=33.33% Model Responders: RS1360780 TT=10.87%

{

Wald (1, 97) = 3.70, p<=.05 High FPS Extinguishers: RS1360780 TT=50.00% Model Responders: RS1360780 TT=16.30%

{

Wald (1, 97) = 3.67, p<=.05

In tru s io n s A v o id a n c e /N u m b in g H y p e ra ro u s a l 1 0 2 0 3 0 P T S D S y m p to m C lu s te rs F e a r P o te n tia te d S ta rtle M o d a l F P S E xtin g u ish e rs H ig h F P S E x tin g u ish e rs H ig h F P S N o n -E x tin g u is h e rs *

SLIDE 31

HPA-Axis Augmentation of Extinction and Recall

Galatzer-levy, et. al. (In Press); Neuropharmacology

n= 127

x t i n c t i

e c a l l 1 2 3 4 A m y g d a la fk b p 5 m R N A H ig h D o se D e x E x tin g u is h e rs N o n -H ig h D o se D e x N o n -E x tin g u ish e rs N o n -H ig h D o se D e x E x tin g u is h e rs * * * * * * * * *

SLIDE 32

Problems in Clinical Trial Research

Variable & multidetermined treatment effects
Personalized medicine problem
Heuristic based clinical constructs
Behavioral phenotyping problem