On page 131 of these notes, a precise formulation of the AdS/CFT correspondence is given by the GKPW dictionary
$$Z_{\text{grav}}[\phi_{0}^{i};\partial M] = \langle \exp \left( - \frac{1}{\hbar} \sum_{i} \int d^{d}x\ \phi_{0}^{i}(x)O^{i}(x) \right) \rangle_{\text{CFT on } \partial M},$$
with
- boundary conditions on the bulk scalars given by
$$\phi^{i}(z,x) = z^{d-\Delta}\phi_{0}^{i}(x) + \text{subleading as } z \to 0,$$ and
- the relation of the mass of the bulk scalar to the scaling dimension of the CFT operator given by
$$m^{2} = \Delta(d-\Delta), \qquad \qquad \Delta = \frac{d}{2} + \sqrt{\frac{d^{2}}{4}+m^{2}\ell^{2}}.$$
Page 132 of the notes then goes on to mention the following:
Similar statements apply to all bulk fields, including the metric, though the boundary condition and formula for the dimension is slightly modified for fields with spin. The boundary conditions on the metric involve a choice of topology as well as the actual metric, which is why we've indicated explicitly that $Z_{\text{grav}}$ depends on the boundary manifold $\partial M$.
I am confused by the final sentence in the above paragraph.
Why are both a choice of topology and a choice of metric needed to write down the boundary conditions on the metric?
Why does this mean that $Z_{\text{grav}}$ depends on the boundary manifold $\partial M$?