I suggest the findings of Zhang, et al., may be explained by levels of dehydroepiandrosterone (DHEA).

It is my hypothesis that schizophrenia is caused by low DHEA in utero. This results in poor brain development. Later in life, cortisol and testosterone act to reduce the effects of already low DHEA and adversely affect brain function as well as maintenance of anatomy. DHEA naturally begins to decline in the early twenties. This is why schizophrenia often occurs in the late teens / early twenties (puberty and loss of DHEA) and is often triggered by a stressful event (cortisol). Therefore, at this time, brain function and maintenance is inhibited and, in the case of cortisol, may be reduced. I suggest this reduces prefrontal function and increases lower brain function, the seat of hallucinations. It is known that DHEA is low in schizophrenia and that DHEA acts positively in neuron growth and function. (DHEA has also been found to be high in schizophrenia. I suggest low DHEA may account for "negative" symptoms and high DHEA may account for "positive" symptoms of schizophrenia.)

It is also my hypothesis that all drugs of abuse first block receptors in the brain which then stimulates "recruitment" of DHEA production to compensate loss of receptors in that part of the brain. Continued use of these drugs ultimately starts a cycle of DHEA stimulation followed by DHEA reduction which becomes a cycle and causes addiction.

So, one smokes, which reduces DHEA, which then rebounds by increasing DHEA. One may say that smoking increases DHEA, eventually. If a non-schizophrenic person, or a schizophrenic person, smokes, the immediate effect may be reduced DHEA. It has been suggested that DHEA "improves cognitive function, and memory enhancement," (J Sex Med. 2011 Nov;8(11):2960-82). Therefore, I suggest smoking may "immediately" reduce DHEA which rebounds and causes differing effects from low DHEA. (This may explain why many who smoke immediately feel relaxed. That is, the reduction in DHEA would reduce immediate stimulation of the brain. Therein, one can "see" one sign of addiction; smoking first relaxes, then stimulates by decreasing, then increasing, DHEA.) So, Zhang, et al., report immediate reductions in cognitive performance in schizophrenic and non-schizophrenic smokers. Smoking should reduce DHEA in schizophrenic and non-schizophrenic smokers. Now, I suggested just above that "negative" symptoms result from low DHEA. Zhang, et al., report that negative symptoms are reduced in schizophrenic smokers. I suggest the ultimate increase in DHEA caused by smoking in schizophrenic patients reduces their negative symptoms.

I suggest the foregoing may explain the findings of Zhang, et al.