Cluster analysis of behavioural and event-related potentials during a contingent negative variation paradigm in remitting-relapsing and benign forms of multiple sclerosis

Table 4 Values represent mean standardized scores for cluster centers after computing k-means cluster analysis

	Cluster 1	Cluster 2	Cluster 3	Cluster 4	Outliers
	CONTROLS = 50%	CONTROLS = 50% RRMS = 11.8% BMS = 10%	RRMS = 68% BMS = 30%	RRMS = 17.4% BMS = 20%	BMS = 20%	p	post-hoc
*Demographics*
Age	28.4 ± 5.7	36.0 ± 9.7	40.7 ± 9.4	38.8 ± 4.1	56 ± 0.0
Dis. duration		5 ± 6.1	6.2 ± 4.6	7.4 ± 3.8	14.1 ± 3.5
EDSS		1.8 ± 1.44	2.1 ± 0.8	1.8 ± 4.7	2.7 ± 0.3
*Behavioural*
RTs	-0.222	0.344	0.908	2,094		<0,001	b c e f
%CRs	0.287	-0.191	-0.632	-3,366		<0,001	c e f
*ERP Latency*
(○) N1 _P4	0.024	0.236	1,817	2,475		<0,001	b c d e
(⊙) P2 _Fz	-0.817	0.877	1,736	2,696		<0,001	a b c d e f
(○) P3 _Fz	-0.556	0,478	2,096	2,921		<0,001	a b c d e
(○) P3 _Cz	-0.652	0.763	2,988	3,897		<0,001	a b c d e f
(○) P3 _Pz	-0.734	0.869	3,133	4,056		<0,001	a b c d e f
(⊙) P3 _Fz	-0.438	0.466	0,863	2,092		<0,001	a b c e f
(⊙) P3 _Cz	-0.579	0.624	1,373	2,350		<0,001	a b c e f
(⊙) P3 _Pz	-0.607	0.582	1,054	2,041		<0,001	a b c e
*ERP amplitude*
eCNV_Cz
eCNV_P4	-0.034	0.052	0,466	0,909		0,038	c
cCNV_Fz	-0.186	0.063	0,850	-0,221		0,047	c
(○) eN1_Fz	0.232	0.019	-0,652	-0,221			d
(⊙) N2_Cz	0.002	0.293	1,370	1,109		0,002	b c d
(⊙) LN_Cz	-0.331	0.043	-0,758	0,368		0,054	-
(⊙) LN _C4	-0.122	0.138	-0,626	0,726		0,018	c d f
	-0.026	0'066	-0,398	-0,028		0,639	-

Significant p values of a new cluster-related ANOVA (using Z-scores as within-subject factors and the different clusters as between-subject factor) are also displayed. The percentage of cases and the contribution of each different experimental group to cluster solutions are also shown. Outlier cases are shown but they were not included in the new ANOVA. In general, ERP latencies had a great weight on splitting the clusters than ERP amplitudes. New demographic values (mean and SD) for age, disease duration and EDSS scale are also shown for each cluster. Differences between clusters for demographic values were not found. Different statistical differences were found between clusters as follows:
[⊙ = target stimuli; ○ = standard stimuli].
a= p < 0.05 between Cluster 1 and 2; b= p < 0.05 between Cluster 1 and 3;
c= p < 0.05 between Cluster 1 and 4; d= p < 0.05 between Cluster 2 and 3;
e= p < 0.05 between Cluster 2 and 4; f= p < 0.05 between Cluster 3 and 4.

ISSN: 1471-2377