Skip to main content
Figure 2 | Gut Pathogens

Figure 2

From: Blood profile holds clues to role of infection in a premonitory state for idiopathic parkinsonism and of gastrointestinal infection in established disease

Figure 2

Distribution of total white cell count, and its decomposition into two Gaussian distributions. Histogram (a) shows loge transformed raw data. Histograms (b) and (c) are generated from parameter estimates obtained by applying a Gaussian mixture model to loge (white cell count). They represent the 'best' mixture to replicate the overall distribution: i.e. two distributions tended to provided a better fit (likelihood ratio test, 3 degrees of freedom, χ2 = 7.25, p = 0.06), (b) representing 80% of data, (c) a shift-to-the-right. There was no evidence that a mixture of Gaussian distributions provided a better fit to distribution of loge transformed lymphocyte (age-corrected) or neutrophil counts (χ2 = 0.74 & 3.49, p = 0.9 & 0.3, respectively).

Back to article page