. 2021 May 19;13:642033. doi: 10.3389/fnagi.2021.642033

Table 3.

Statistics as per feature set and language.

	Semantic features
			English						French
Feature	r_PB	m_HC	m_AD	χ²	p	p_corr.	r_PB	m_HC	m_AD	χ²	p	p_corr.
keyword_to_non_keyword_ratio	−0.43	0.16	0.11	19.3	***	***	−0.42	0.15	0.11	8.1	**	0.10
max_word_frequency_IU	−0.27	0.0003	0.0003	7.5	***	0.13	−0.34	0.0003	0.0002	5.3	*	0.47
mean_word_frequency_all	−0.37	0.0089	0.0072	14.7	***	***	−0.38	0.0075	0.0066	6.8	**	0.20
num_unique_IU	−0.60	10.94	6.87	37.8	***	***	−0.64	10.24	5.50	18.6	***	***
num_unique_keywords	−0.60	11.83	7.26	37.3	***	***	−0.60	11.40	5.73	16.6	***	**
percentage_of_keywords_mentioned	−0.60	0.10	0.06	37.3	***	***	−0.60	0.07	0.04	16.6	***	**
total_IU_density	−0.42	0.14	0.10	18.7	***	***	−0.36	0.14	0.11	5.9	**	0.34
total_IU_efficiency	−0.54	0.26	0.16	30.2	***	***	−0.46	0.34	0.21	9.8	**	*
num_total_keywords	−0.43	15.42	10.46	19.3	***	***	−0.57	13.44	6.41	15.1	***	**
unique_IU efficiency	−0.56	0.18	0.10	32.4	***	***	−0.40	0.27	0.17	7.2	**	0.16
unique_IU ratio	−0.45	0.10	0.07	21.2	***	***	−0.34	0.11	0.09	5.5	*	0.43
	Syntactic features
	English						French
Feature	r_PB	m_HC	m_AD	χ²	p	p_corr.	r_PB	m_HC	m_AD	χ²	p	p_corr.
ADP_count	−0.28	7.83	5.39	8.0	**	0.20	−0.50	14.44	6.95	11.4	***	*
ADP_ratio	−0.34	0.06	0.05	12.3	***	*	−0.51	0.13	0.09	11.9	***	*
AUX_ratio	−0.35	0.10	0.09	12.7	***	*	0.30	0.04	0.06	4.1	*	1.00
DET_count	−0.26	17.35	13.52	7.3	**	0.31	−0.45	17.72	9.95	9.4	**	0.09
DET_ratio	−0.43	0.15	0.12	19.3	***	***	−0.32	0.17	0.14	4.7	*	1.00
NOUN_count	−0.34	21.12	15.59	11.9	***	*	−0.49	20.76	11.09	11.0	***	*
NOUN_ratio	−0.48	0.18	0.14	23.8	***	***	−0.38	0.19	0.15	6.7	**	0.42
PRON_ratio	0.25	0.07	0.09	6.4	*	0.51	0.51	0.11	0.18	12.1	***	*
PUNCT_count	0.21	15.15	20.69	4.7	*	1.00	−0.38	1.06	0.32	6.5	*	0.47
PUNCT_ratio	0.36	0.13	0.18	13.8	***	**	−0.34	0.01	0.00	5.4	*	0.91
	Paralinguistic features
	English						French
Feature	r_PB	m_HC	m_AD	χ²	p	p_corr.	r_PB	m_HC	m_AD	χ²	p	p_corr.
bandwidth_mean	0.22	2,022.85	2,153.46	5.0	*	1.00	0.32	2,176.25	2,323.75	4.8	*	1.00
energy_skewness	0.23	0.17	0.32	5.4	*	1.00	0.47	-0.26	0.49	10.4	**	0.26
mfcc1_mean	−0.20	-1.87	-4.43	4.4	*	1.00	−0.36	-5.14	-7.87	6.0	*	1.00
mfcc1_skewness	0.23	0.19	0.48	5.8	*	1.00	0.44	-0.31	0.12	8.8	**	0.63
mfcc10_kurtosis	0.23	0.73	1.00	5.5	*	1.00	0.37	0.40	0.67	6.4	*	1.00
mfcc4_kurtosis	0.22	0.92	1.47	4.9	*	1.00	0.41	0.51	1.02	7.7	**	1.00
normalized_loudness_std	−0.34	0.20	0.18	11.8	***	0.12	−0.56	0.23	0.20	14.4	***	*
ratio_speaking	−0.27	0.46	0.37	7.6	**	1.00	−0.57	0.64	0.48	15.2	***	*
speech_rate	−0.28	1.92	1.41	8.5	**	0.73	−0.47	3.12	2.32	10.2	***	0.28

Point-biserial correlation coefficient r_PB, correlating each feature with the nominal group variable (AD, HC), feature means for HC and AD, χ² value of the non-parametric Kruskal–Wallis H-test for group differences between AD and HC, p-value and Bonferroni-corrected p-value. Significances: *** 0.001, ** 0.01, * 0.05. Feature names in Bold indicate that they are significant after Bonferroni-correction in both languages.