I’ve been thinking some more about the recent Woodley et al paper that respected blogger Dr. James Thompson has been writing about. The paper implies that since the 19th century or so, brain weight has been increasing by about 0.21 standard deviations (SD) per half century (if you average the rate of brain size gains in the samples reported in table 2 of the paper).  In a previous post, I argued that true gains reported in the study are more like 0.29 SD per half century.

However even 0.29 SD per half century sounds way too low.  For example, scholar Richard Lynn notes that head circumference in British one-year-olds and British seven-year-olds went up 1.5 cm and 2 cm in 50 years.  Given that the SD for head circumference among whites at both ages is 1.5 cm (see table 1 of this paper), that’s an increase of 1.16 SD per half-century.

I realize that gains in children may not match gains in adults,  and that the gains might not be perfectly linear throughout the 20th century, but given that most brain growth has occurred by age seven, and that head circumference in children is very highly correlated with brain size, it seems strange the brain size has increased so little in adults when head circumference in in kids has increased so much.

Pumpkin Person’s analysis

I decided to look at the oldest representative sample of age controlled white brain weight in a Western country I could find and compare it to the newest representative sample of age controlled white brain weight in a Western country I could find.  Because  I don’t know the age controlled standard deviation for brain weight and non-age controlled SDs are not ideal in my opinion, I decided to convert autopsied brain weight to in vivo volume (following Rushton 1997) and use the inferred standard deviations of Caucasian enlisted army men and women (91 cm3 and 90cm3 respectively, see table 1 of Rushton’s data ) since skull measurements, especially in young adults are much less age sensitive than brain weights in mixed-age adults.

British 48.5 year-olds, born in 1860, dead in 1908.5

A study by Miller and Corsellis found that British men and women born in 1860 had autopsied brain weights of 1372 g and 1242 g respectively, which following Rushton’s procedure of dividing by 1.09 and multiplying by 1.036, converts to in vivo brain volumes of 1304 cmand 1180 cm3  respectively.  It’s unclear how old they were when they died but given that the data-set included brains from no earlier than 1907, and no older than 50, I think it can be inferred that they died at 47-50 in the years 1907-1910 (let’s say 48.5 and 1908.5 for simplicity).

White American 48.5 year-olds, born in 1931.5, dead in 1980

Meanwhile a study by Ho et al (1980) found that White American men and women (who presumably died circa 1980) had autopsied brain weights of about 1392 g and 1252 g at age 60 and 1570 g and 1339 g at age 25 respectively.  Assuming age differences in brain weight are linear, we can deduce that 48.5 year white American men and women dead circa 1980 had autopsied brain weights of 1450 g (1378 cm3 in vivo) and  1281 g  (1218 cm in vivo) respectively.

Secular increase in brain size: 0.44 SD per half-century

men women
brain size of british 48.5 year olds dead in 1908.5 1304 cm3 1180 cm3
brain size of white american 48.5 year olds dead in 1980 1378 cm3 1218 cm3
absolute difference in 71.5 years 74 cm3 38 cm3
standard deviation (sd) 91 cm3 90 cm3
difference in sd units per half century 0.57 0.3

Averaging men and women together, brain size has increased by 0.44 SD per half century, roughly twice the gains implied by Woodley et al paper.  However my analysis is somewhat ad hoc and did not include as many data points as the  Woodley et al paper.

White adult crania has been increasing at a rate of 0.66 SD per half-century, predicting a 0.44 SD per half-century increase in brain size

There’s also data showing that white male crania being 200 cm3 bigger in white American men born in 1980 compared to white American men born in 1820. Similarly, white female crania has increased 180cmover the same time period.  Given that brain volume has an SD of 91 cm3  in men and 90 cm3 in women cited above, that’s an increase of about 2.2 SD for men and 2 SD for women, so let’s split the difference and say it’s 2.1 SD in 160 years, or 0.66 SD per half-century.  However this data is presumably from external head measurements, not direct measures of endocranial volume or brain size.   What is the correlation between head size and brain size?  Scholars Rushton and Ankney write:

The evidence shows that external head size is a good proxy for brain volume. Head perimeter correlates with brain mass at autopsy from birth through childhood at correlation values of 0.80 to 0.98 (Brandt, 1978; Bray, Shields, Wolcott, & Madsen, 1969; Cooke, Lucas, 132 Chapter 6 Yudkin, & Pryse-Davies, 1977). It correlates with MRI brain volume at an average value of 0.66, based on fjve studies (0.55 in 10 pairs of identical twins aged 24-43 years, Tramo et al., 1998; 0.66 in 34 pairs of brothers aged 20-35 years, Wickett et al., 2000; 0. 74 in 103 university students of both sexes in Turkey, Tan et al., 1999; 0.56 in 83 normal controls aged 8-46 years in the United States, Aylward, Minshew, Field, Sparks, & Singh, 2002; and 0.79 in 96 high school graduates of both sexes in Chile, lvanovic et al., 2004).

Given that the studies of almost all adults average a 0.66 correlation between external head size and brain size, then a 0.66 SD increase in adult head size may predict a 0.66(0.66) = 0.44 SD increase in brain size which is precisely what my brain size analysis found.  This doesn’t prove that my analysis is correct, of course, but it might be suggestive.