Hi guys, i´m new kind of new in the multilevel modeling in brms world, and actually im litte confuse.

The point is to fit this model in brms

y_i = \beta_1 x_{1i}+ \beta_2 x_{2i}+\epsilon_i

in my understanding this is the way

```
brm(y~-1+x1+x2,data = Data)
```

The above is the fixed effects model which is consistent with other frequentist libraries and I have no problem with it.

the confusion comes with the random effects model for which I search for values of \beta_j

present variation with respect to another variable, let’s say “Study”, so with a little search in the brms documentation I found that I could model the random effects so that the slopes vary across “ Study ” by

```
brm(y~-1+x1+x2+(x1+x2|Study),data=Data)
```

I don’t know exactly if this is 100% correct although in the same way I find consistent results with respect to other libraries.

But the real point of my post is that I do not yet find, if it exists, a way to model the correlation between the random effects in a way other than by the prior distributions and `cosy`

or `cor_fixed`

but I did not clearly understand how they are implemented

i,e say \theta_1 and \theta_2 the true random effects, the intention is to model Cor(\theta_1,\theta_2) =\frac{1}{2}

Any help or some kind of guidance with brms will be a great help

THANK YOU