Are there specific disabilities of number processing in adolescent patients with Anorexia nervosa? Evidence from clinical and neuropsychological data when compared to morphometric measures from magnetic resonance imaging.

The cerebral effect of the loss of body weight in Anorexia nervosa (A.n.)--the so-called 'pseudoatrophy'-- is well known and confirmed by several neuroimaging studies. Another subject of intensive research has been whether A.n. leads to specific cognitive impairments, especially of intelligence. However, there are no previous studies on the relations between the cerebral changes, intelligence performance, and disorders of number processing in adolescent patients with A.n. We examined n = 18 inpatients with A.n. (means at admission: age 14.5 years, SD 1.59; BMI 14.9, SD 1.36), diagnosed according to ICD-10 criteria at three different timepoints: at admission to treatment (T1), with 50 % restoration of their normal weight (T2), and with normal weight (T3). At each timepoint, a cerebral MRI scan was obtained. Based on the MRI we determined the volume of the external and internal cerebrospinal fluid cavities, fissures of Sylvius, the surface of mesencephalon and pons, and surface and length of the Corpus callosum. At T1 and T3, a neuropsychological examination was conducted including tests of the general fluid ability and general cristallized ability of intelligence (CFT-20), as well as tests of vocabulary and number processing. The same instruments were given to a group of matched controls (means: age 15.8 years, SD 1.57; BMI 20.5, SD 2.3) at one timepoint. We could show a significant volume difference of the lateral ventricles and the fissure of Sylvius between patients at T, and controls, which abaded with the patient's weight restoration. But a significant surface deficit of the mesencephalon, and less pronounced in the pons, persisted to T3 in patients when compared to controls, suggesting a selectivity of the cerebral changes in A.n. The neuropsychological examinations revealed significant changes in test performance for both the general intelligence test and number processing. At T1 the number processing performance was significantly lower in patients when compared to controls. However, when the patients had restored their normal body weight, we found 2.02 % with a 'severe disorder of arithmetic skills' and 4.45 % with a 'functional disorder of arithmetic skills'. This combined prevalence of 6.47 % of patients with a subnormal arithmetic performance is analogous to that in the normal population.