The arrows indicate the way in which two concepts are associated. Causaly reads and understands text similarly to how humans do and represents the directionality of relationships between 2 Concepts with an arrow. There are four different types of arrows:
A. Upregulate arrow 
The entities are moving in the same direction:
i. More of Concept A results in more of Concept B, e.g. A drug induces a side effect.
ii. Less of Concept A results in less of Concept B, e.g. inhibition of a target reduces the progression of the disease.
B. Downregulate arrow 
Two entities are moving in the opposite direction:
i. More of Concept A results in less of Concept B, e.g. a drug is a treatment for a disease.
ii. Less of Concept A results in more of B, e.g. inhibition of a target induces progression of the disease.
C. Unidirectional arrow
There is an effect of A on B but the language does not mention an increase or decrease. For example, when a target plays a role/has an effect/modulates a disease.
D. Bidirectional arrow 
Two entities are associated or correlated but the directionality is not clear. For example, an interaction between two targets.
E. Negating arrows 
Refuting statements are also captured in Causaly and are depicted with a strikethrough arrow for all 4 arrow types.
