Coherent obsessional experiments for linear logic proof-nets