Limits...
Making Large-Scale Networks from fMRI Data.

Schmittmann VD, Jahfari S, Borsboom D, Savi AO, Waldorp LJ - PLoS ONE (2015)

Bottom Line: However, this approach generally results in a poor representation of the true underlying network.As a result, pairwise correlation networks can lead to fallacious conclusions; for example, one may conclude that a network is a small-world when it is not.We conclude that using partial correlations, informed by a sparseness penalty, results in more accurate networks and corresponding metrics than pairwise correlation networks.

View Article: PubMed Central - PubMed

Affiliation: Department of Methodology and Statistics/Social and Behavioral Sciences, Tilburg University, Tilburg, the Netherlands.

ABSTRACT
Pairwise correlations are currently a popular way to estimate a large-scale network (> 1000 nodes) from functional magnetic resonance imaging data. However, this approach generally results in a poor representation of the true underlying network. The reason is that pairwise correlations cannot distinguish between direct and indirect connectivity. As a result, pairwise correlation networks can lead to fallacious conclusions; for example, one may conclude that a network is a small-world when it is not. In a simulation study and an application to resting-state fMRI data, we compare the performance of pairwise correlations in large-scale networks (2000 nodes) against three other methods that are designed to filter out indirect connections. Recovery methods are evaluated in four simulated network topologies (small world or not, scale-free or not) in scenarios where the number of observations is very small compared to the number of nodes. Simulations clearly show that pairwise correlation networks are fragmented into separate unconnected components with excessive connectedness within components. This often leads to erroneous estimates of network metrics, like small-world structures or low betweenness centrality, and produces too many low-degree nodes. We conclude that using partial correlations, informed by a sparseness penalty, results in more accurate networks and corresponding metrics than pairwise correlation networks. However, even with these methods, the presence of hubs in the generating network can be problematic if the number of observations is too small. Additionally, we show for resting-state fMRI that partial correlations are more robust than correlations to different parcellation sets and to different lengths of time-series.

No MeSH data available.


Related in: MedlinePlus

Recovery results of the four estimation methods for additional random network  (with the high density of 3%).True −− network metrics indicated where appropriate. The dotted line ⋯ shows the level of the false positive rate, above which the absolute number of false positive edges even exceeds the absolute number of edges in the true network.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4556712&req=5

pone.0129074.g015: Recovery results of the four estimation methods for additional random network (with the high density of 3%).True −− network metrics indicated where appropriate. The dotted line ⋯ shows the level of the false positive rate, above which the absolute number of false positive edges even exceeds the absolute number of edges in the true network.

Mentions: As shown above, in the two networks with hubs, all methods perform worse. In these networks, the maximum degree is much larger than in the networks without hubs (see Table 1). However, these two networks also have a larger number of edges (density of 3%), in order to make a network with large-degree nodes and still be connected, than the networks without hubs (density of 0.3%). To separate the effects of density and hubs, we analyzed the complementary random network (i.e., without hubs) with a density of 3% (). The results support the hypothesis that the presence of hubs causes the decrease in perfomance, rather than the lower density of the network. The true positive and false positive rates of network (Fig 15) show much better performance of the partial correlation networks than the pairwise correlation networks with hubs ( and SW-H, Figs 10 and 11), but also, slightly worse performance than in the sparser random network .


Making Large-Scale Networks from fMRI Data.

Schmittmann VD, Jahfari S, Borsboom D, Savi AO, Waldorp LJ - PLoS ONE (2015)

Recovery results of the four estimation methods for additional random network  (with the high density of 3%).True −− network metrics indicated where appropriate. The dotted line ⋯ shows the level of the false positive rate, above which the absolute number of false positive edges even exceeds the absolute number of edges in the true network.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4556712&req=5

pone.0129074.g015: Recovery results of the four estimation methods for additional random network (with the high density of 3%).True −− network metrics indicated where appropriate. The dotted line ⋯ shows the level of the false positive rate, above which the absolute number of false positive edges even exceeds the absolute number of edges in the true network.
Mentions: As shown above, in the two networks with hubs, all methods perform worse. In these networks, the maximum degree is much larger than in the networks without hubs (see Table 1). However, these two networks also have a larger number of edges (density of 3%), in order to make a network with large-degree nodes and still be connected, than the networks without hubs (density of 0.3%). To separate the effects of density and hubs, we analyzed the complementary random network (i.e., without hubs) with a density of 3% (). The results support the hypothesis that the presence of hubs causes the decrease in perfomance, rather than the lower density of the network. The true positive and false positive rates of network (Fig 15) show much better performance of the partial correlation networks than the pairwise correlation networks with hubs ( and SW-H, Figs 10 and 11), but also, slightly worse performance than in the sparser random network .

Bottom Line: However, this approach generally results in a poor representation of the true underlying network.As a result, pairwise correlation networks can lead to fallacious conclusions; for example, one may conclude that a network is a small-world when it is not.We conclude that using partial correlations, informed by a sparseness penalty, results in more accurate networks and corresponding metrics than pairwise correlation networks.

View Article: PubMed Central - PubMed

Affiliation: Department of Methodology and Statistics/Social and Behavioral Sciences, Tilburg University, Tilburg, the Netherlands.

ABSTRACT
Pairwise correlations are currently a popular way to estimate a large-scale network (> 1000 nodes) from functional magnetic resonance imaging data. However, this approach generally results in a poor representation of the true underlying network. The reason is that pairwise correlations cannot distinguish between direct and indirect connectivity. As a result, pairwise correlation networks can lead to fallacious conclusions; for example, one may conclude that a network is a small-world when it is not. In a simulation study and an application to resting-state fMRI data, we compare the performance of pairwise correlations in large-scale networks (2000 nodes) against three other methods that are designed to filter out indirect connections. Recovery methods are evaluated in four simulated network topologies (small world or not, scale-free or not) in scenarios where the number of observations is very small compared to the number of nodes. Simulations clearly show that pairwise correlation networks are fragmented into separate unconnected components with excessive connectedness within components. This often leads to erroneous estimates of network metrics, like small-world structures or low betweenness centrality, and produces too many low-degree nodes. We conclude that using partial correlations, informed by a sparseness penalty, results in more accurate networks and corresponding metrics than pairwise correlation networks. However, even with these methods, the presence of hubs in the generating network can be problematic if the number of observations is too small. Additionally, we show for resting-state fMRI that partial correlations are more robust than correlations to different parcellation sets and to different lengths of time-series.

No MeSH data available.


Related in: MedlinePlus