Encoding of Race Categories by Single Neurons in the Human Brain
Round 1
Reviewer 1 Report
In their manuscript, Valdez and colleagues use data taken in patients suffering from drug-resistant epilepsy. In preparation for surgical treatment, microwires were implanted to the brains of the patients to localize seizure onset zones. In the study presented here, the implanted microwires were used to determine neuronal activity in the hippocampus, the prefrontal cortex, the anterior cingulate cortex, and the amygdala in response to the screen presentation of “synthetic faces” differing in color of skin. Thereby, the authors claim to show the “encoding of race categories by single neurons in the human brain”. The manuscript is well written. However, I have methodological and content concerns.
Major concerns:
(1) Skin color and ethnicity are not the same thing. The authors show – if at all (see further concerns) – that there could be different responses in the brain areas tested depending on skin color, but not for ethnicity. From the results presented, I cannot see a “race categorization”.
(2) You can only prove that there is a skin color categorizing in specific brain areas when responses in the brain areas of interest are similar (not significantly different) for in-group response and for out-group response in experimental groups differing in skin color. Otherwise, you cannot exclude the possibility that the responses are not due to similarity between presented and own skin color, but for any other reason. In other words: You have to examine the responses of at least a second group, which differs in skin color from the group examined in your manuscript. From your results taken in a single skin color group, you cannot claim to show the “encoding of race categories by single neurons in the human brain”.
(3) The synthetic faces were always presented in front of a white background. How can you exclude that you actually determined neuronal response to ethnicity and not, for instance, to contrast or change in screen brightness during presentation? Maybe you would get similar results presenting pictures of black, brown and white rabbits…
(4) The amygdala consists of many individual nuclei that process different information and differ in function. The placement of the microwires, in view of this, is very rough. Why do you think you have received individual results that you can compare. Actually, neurons could have been measured in very different nuclei.
(5) Line 103: Repeated sessions were treated as independent samples? In my opinion, this is pseudoreplication and a serious methodological issue.
Line 156: What was the sampling rate during measurement? This is important!
(6) Some of the figures (and figure legends) and tables are hard to understand. For instance, what does the label of the y-axis of Figure 3 mean? What do you mean by saying: “Percentages of significant neurons (SUA) for stimulus race category” What are “significant neurons”? In Table 1: Percentage of what? All neurons in the respective brain area? Neurons tracked? Some of the information can be find in the main text. However, it would be so much easier to read your manuscript, when needed information would be available at place.
Minor concerns:
(1) Data of African-American patients were excluded? But they showed no effects to the participant’s race? What do you mean?
(2) Figure 1: The order in the legend does not match the order in the figure. “From left to right: neutral, positive, negative…” would be correct.
(3) Line 208: Why did the recording period differ in the range form 200 to 1000 ms? Why did you not use a fixed recording period?
(4) Line 289 to 296: A second legend to Figure 3?
(5) Line 297: The figure is huge. Why is this necessary?
(6) Line 302: Table 1 can only be understood after reading the main text. Maybe the authors could present results in the table in a way making this obsolete?
(7) Line 330: What do you mean with the term “event waveform”? I have not understood what you show in Fig. 4c.
Author Response
Please see attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Title Encoding of Race Categories by Single Neurons in the Human Brain
Manuscript ID neurosci-1774867
This is a very complex work, but well designed, involving a topic that is somewhat controversial due to its biopsychosocial implications, but on the other hand of interest in advanced neurology and single neurons studies.
It is important to highlight the distinctions between the concepts of race and ethnicity that involve the erroneous notion of biological distinction and the cultural differentiation of diverse groups. The concept of race is still widely used by the vast majority of people, especially when referring to the skin color. However, the idea that there are biologically distinct races among humans is a mistake. Scientific theories of the 18th and early 19th centuries, such as those of Joseph Arthur de Gobineau (1816–1882), were dedicated to the creation of methods to differentiate races between human groups through, mainly, phenotypic traits (skin color , hair, skull shape). A large part of these theoretical works used as a justification the “higher” degree of development of European countries, which enriched themselves thanks to the exploitation of “inferior” civilizations. However, the “racial sciences” lost strength and became completely discredited by the scientific community after the Second World War. Later work, such as the human genome mapping project, proved that genetic distinctions between different human groups, which have different physical characteristics, are not sufficient to justify the separation into races. The most diverse researches have shown that the genetic variation that exists between geographically separated groups is very close to the genetic variation between subjects of the same group. However, even if the biological concept of race is wrong, the social impact still exists. Social phenomena that are based on the notion of racial differentiation because of skin color, for example, are still strongly rooted in our communities. In this sense, the concept of race is still widely used by Sociology to understand the different relationships that are established around the evaluative notion that exists regarding color. Racial distinctions go beyond human differentiation due to physiological characteristics, they are inserted in the reproduction of inequalities in our social environment.
If the concept of race is linked to the biological distinction of human groups based on physical differences, the concept of ethnicity is associated with the cultural practices and constructions that differentiate one group from the others. Ethnic groups are distinguished by cultural characteristics, such as language, religion, clothing, among other aspects. Ethnicity is an essentially social phenomenon, as it is a continuous process of cultural transmission between different generations based on contact and participation in the social environment in which ethnicity is configured. There are many people who see their ethnicity as a huge part of their individual identity. It is more common to see examples of this phenomenon in immigrant communities that settle in another country. The maintenance of traditions or celebrations that exalt cultural characteristics of the group serves as a way of maintaining an ethnic identity that, while differentiating individuals from other groups, also unites them through the familiarity with the cultural traits they share.
Having placed these considerations, the following additional questions remain:
1-I understand the difficulty of obtaining patients for this study, but how was the size of this sample statistically selected? Add in text. This seems small for large conclusions (only 14 patients). Moreover, the results of 1 patient were excluded due to technical errors and there were practically only Caucasian participants with epilepsy who may have their cognition largely affected.
2-Could this have been compensated by viewing 120 faces per session? How was this amount of face viewing per session statistically selected? What would be the ideal minimum size considering that in this study there was no robust African-American sample, most right Handedness, different brain areas with seizure foci? The patients in this experiment have epilepsy, which may broadly affect their cognition.
3-Women and men have differences in the perception of race? Children, adults and seniors? What does the literature talk about it? Add.
4-Was free informed consent obtained from the patients? Where and what is the registration number for conducting this research? Add.
5-Were the patients under the effects of medication during data collection in this cognitive experiment, since they were hospitalized? Explain these details in the text.
6-It would be interesting to consider including a schematic drawing or photo of the implantation sites chosen for the electrodes or of a case with the wires implanted stereotatically.
7-Were all the microwire bundles in the target structures? How was it confirmed?
8-Regarding the slight bias to interpreting neutral expressions as negative, what was the percentage of error for neutral expressions of black races identified as negative faces? Was this number higher in white patients or in black and Hispanic patients combined?
I am grateful for the opportunity to contribute to the review of this study and hope to follow up on future results. I suggest inclusion of combination monitoring of the patients' face with other methods, such as infrared thermal imaging (peripheral response assessment).
Author Response
Please see attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Thank you for clarifying many of the aspects mentioned before. The authors were able to significantly improve the manuscript in many sections. Nevertheless, there are still concerns to address.
Major concerns:
(1) Skin color and “race”: The authors answered that “skin color was the salient, “driving” feature that supports such accurate responding, with some potential contributions from other covarying features.” Why then not saying that single neurons in the human brain encode skin color (for instance in the manuscript title)? Why saying that they encode race? For me, it is hard to accept that the authors equate skin color and race. Yes, they used synthetic faces. That seems to be the standard in this kind of research. On the other hand, if I think about my own environment, skin color would not be at all suitable to determine ethnicity.
(2) Reviewer’s concern: Line 103: Repeated sessions were treated as independent samples? In my opinion, this is pseudoreplication and a serious methodological issue.
Author’s response: We appreciate the reviewer's concern with statistical accuracy. There are two reasons that this is not considered pseudoreplication in this type of recording. Firstly, the depth electrodes and microwires extend approximately 5 cm from their point of fixation at the skull. Given this and that the patient is normally moving around the room between experimental sessions to go to the bathroom, etc., as well as potentially having seizures, it is very unlikely that the same neuron is being recorded across different sessions.
Reviewer’s Response: Could you give references for that in your manuscript? How much does the the brain tissue move per hour/day? In this context: Are you sure that the neurons you recorded from are the same within a session?
Secondly, one would need to posit that there is a difference in the likelihood of recording from the same neuron depending on whether the neuron is responding to race in order for this to have an effect on the estimated fraction of the neurons showing a response to race. The following example shows why this is true: one has recorded 100 neurons and 30 of them have a response to race. If one then duplicates all of them, the new estimated fraction of neurons with a response to race is 60 / 200 or again 30%. The existence of such a differential likelihood of recording the same neuron in a second session dependent on the response to race seems a remote possibility.
Reviewer’s Response: The argument doesn't make sense to me. If the brain shifts and wires change positions, then the same neurons could still be recorded to a large extent. This depends on the extent of shift. Doesn’t it? On the one hand, you say that the tissue is significantly shifting. In consequence, you say that you had have recorded from completely different neurons in each of the sessions. On the other hand, you say that you measured from specific structures of the brain. Which should apply now? Could you present data indicating the extent of tissue shift? You measure several times, but a certain neuronal fraction could actually be recorded repeatedly. To me, that is not independent sampling. Perhaps a multivariate analysis factoring in the individual patient would be more appropriate here.
Minor concerns:
(1) Reviewer’s concern: Line 297: The figure is huge. Why is this necessary?
Author’s response: We assume the reviewer is referring to Figure 4. This figure shows actual peri-stimulus time histograms of the firing of a neuron in response to the presentation of images from the different race categories. This type of figure is normally shown in papers describing this type of human single neuron recording so that readers can visually judge the strength of responses to the different categories. While we tend to think the statistical analyses are more meaningful, given the long history of such presentation, we have retained this presentation as suggested by a prior reviewer for consistency.
(2) Line 169: “an a”? Delete “an”.
Why figure 4? I meant Figure 3, for example. Why is it so big?
(3) Line 364: The graph has slipped and is partially obscured by Table 2.
(4) Line 566: What is the scaling?
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 3
Reviewer 1 Report
The authors were able to address most of my concerns and to improve the text and the descriptions. I can understand the authors' explanations regarding the use of individual sessions as independent measurements. I still disagree, though. However, the explanation in the manuscript is suitable for readers to decide for themselves whether or not they find the approach and analysis suitable. Therefore I do not have further concerns against publishing the manuscript in its current form.
