Skip to main content

Table 2 Final diagnoses of stroke mimics

From: Use of deep artificial neural networks to identify stroke during triage via subtle changes in circulating cell counts

 

Total sample (n = 116)

Training set (n = 87)

Test set (n = 29)

Seizure n (%)

20 (17.2)

16 (18.4)

4 (13.8)

Complex migraine n (%)

13 (11.2)

8 (9.2)

5 (17.2)

Hypertensive encephalopathy n (%)

7 (6.0)

5 (5.7)

2 (6.9)

Psychogenic event / psychiatric disorder n (%)

7 (6.0)

6 (6.9)

1 (3.4)

Hypotension / syncope n (%)

7 (6.0)

5 (5.7)

2 (6.9)

Sepsis / infection n (%)

7 (6.0)

6 (6.9)

1 (3.4)

Adverse drug reaction n (%)

6 (5.2)

4 (4.6)

2 (6.9)

Space occupying lesion n (%)

6 (5.2)

4 (4.6)

2 (6.9)

Bell's palsy n (%)

5 (4.3)

4 (4.6)

1 (3.4)

Alcohol intoxication n (%)

4 (3.4)

2 (2.3)

2 (6.9)

Hypoglycemia n (%)

3 (2.6)

2 (2.3)

1 (3.4)

Dementia n (%)

3 (2.6)

3 (3.4)

0 (0.0)

Dehydration n (%)

3 (2.6)

3 (3.4)

0 (0.0)

Vestibular dysfunction n (%)

3 (2.6)

1 (1.1)

2 (6.9)

Peripheral neuropathy n (%)

2 (1.7)

2 (2.3)

0 (0.0)

Transient global amnesia n (%)

2 (1.7)

1 (1.1)

1 (3.4)

Parkinson’s disease n (%)

2 (1.7)

2 (2.3)

0 (0.0)

Multiple sclerosis n (%)

2 (1.7)

2 (2.3)

0 (0.0)

Hepatic encephalopathy n (%)

2 (1.7)

1 (1.1)

1 (3.4)

Other n (%)

12 (10.3)

10 (11.5)

2 (6.9)