Auditory white matter tract development in infants exposed to HIV and antiretrovirals

This study suggests that HIV and antiretroviral exposure may cause subtle, uncorrected delays in auditory white matter tract maturation in infants, particularly affecting those exposed post-conception and potentially impacting language processing, although these findings did not reach statistical significance after multiple comparison correction.

The Gist

🎧 The Big Question: Does "HIV Exposure" Change a Baby's Brain Wiring?

Imagine a baby's brain as a massive, bustling city. The white matter tracts are the highways and fiber-optic cables that connect different neighborhoods (like the hearing center and the language center). For a baby to hear clearly and eventually learn to speak, these highways need to be smooth, well-paved, and fast.

This study looked at three groups of babies in South Africa:

1. The "Unexposed" Group: Babies whose mothers did not have HIV.
2. The "Early Treatment" Group: Babies whose mothers had HIV but started medication before getting pregnant.
3. The "Late Treatment" Group: Babies whose mothers had HIV but started medication after getting pregnant.

The researchers wanted to know: Does being exposed to HIV or the medication in the womb change how these "highways" are built? And if they are built differently, does it affect how the baby learns to talk later on?

---

🔍 How They Looked Inside the Brain

You can't just look at a baby's brain with a flashlight. The researchers used a special type of MRI scan called DTI (Diffusion Tensor Imaging).

Think of DTI like a traffic camera system. Instead of taking a photo of the cars, it tracks how water molecules move through the brain's "roads."

• FA (Fractional Anisotropy): Think of this as the smoothness of the road. High FA means a smooth, paved highway where traffic (signals) moves fast. Low FA means a bumpy, pothole-ridden road.
• MD (Mean Diffusivity): Think of this as the traffic density. High MD means the road is loose and disorganized (like a dirt path), while low MD means it's packed and organized (like a concrete highway).

They scanned the babies when they were just 2 to 5 weeks old (while they were sleeping naturally, no drugs needed!) and then tested their language skills when they were about 1 year old.

---

📉 What They Found: The "Bumpy Road" Effect

1. The Roads Were Slightly Rougher
The researchers found that the babies exposed to HIV (even if they weren't infected themselves) had some "bumpier" roads in their hearing centers compared to the unexposed babies.

• The "Late Treatment" Group: The babies whose moms started meds later in pregnancy showed the most "bumpy" roads (lower smoothness) in the connections between the hearing center and the rest of the brain.
• The "Left Side" Issue: The problems were mostly on the left side of the brain, specifically in a tiny relay station called the Medial Geniculate Nucleus (MGN). You can think of the MGN as the switchboard operator that takes sound signals from the ear and sends them to the brain's language center.

2. The "Late Treatment" vs. "Early Treatment" Mystery

• Early Treatment: Even though these moms took meds for a longer time, their babies still showed some signs of "looser" roads (higher traffic density/MD).
• Late Treatment: These babies showed the most "bumpy" roads (lower smoothness/FA).

The Takeaway: It seems that HIV exposure itself (perhaps due to the mother's immune system being stressed) might be the main culprit causing these delays, rather than the length of time the baby was on medication. It's like a construction crew working on a highway; even if they start early, the fact that the ground was unstable (HIV exposure) might still cause some cracks.

---

🗣️ The Twist: The Babies Are Still Talking Fine!

Here is the most surprising part: Despite having slightly bumpier roads, the babies' language scores were normal.

When the researchers tested the babies at 9–14 months old, the "HIV-exposed" babies spoke and understood just as well as the "unexposed" babies.

So, what happened?

• The "Unexposed" Group: In these babies, the smoother the roads, the better they spoke. It's a direct link: Better highway = Better driver.
• The "Exposed" Group: In these babies, the link was broken. Even if their roads were bumpy, they still drove well.

The Metaphor:
Imagine two drivers.

• Driver A (Unexposed): Needs a perfect highway to drive fast. If the road is bumpy, they slow down.
• Driver B (Exposed): Has a slightly bumpy highway, but they have learned to drive differently. Maybe they take a different route, or they drive more carefully. They aren't slower; they just got there in a different way.

This suggests that the HIV-exposed babies' brains might be rewiring themselves or finding alternative paths to learn language, even if the "main highway" looks a bit different on the scan.

---

🏁 The Bottom Line

1. HIV exposure changes the blueprint: Being exposed to HIV or antiretroviral drugs in the womb seems to slightly delay the "paving" of the brain's hearing highways, especially on the left side.
2. Timing matters less than the exposure: Whether the mom started meds early or late, the exposure itself seems to be the main factor, though starting meds later might make the roads a bit bumpier.
3. The brain is resilient: Even with these "bumpy roads," the babies are developing language skills normally. Their brains are finding creative ways to compensate.

The Future:
The researchers say we need to keep watching these children. Just because they are fine at 1 year old doesn't mean they won't face challenges later as the brain gets more complex. It's like checking a car's engine before a long road trip; the engine might run fine now, but we want to make sure it holds up for the whole journey.

In short: The "construction" of the brain's hearing wires was slightly delayed, but the "drivers" (the babies) are still navigating the city successfully.

Technical Summary

1. Problem Statement

While it is established that perinatal HIV infection and exposure to antiretroviral therapy (ART) can affect brain development and language outcomes, the specific impact on the central auditory system in HIV-exposed uninfected (iHEU) infants remains unclear. Previous studies have yielded inconsistent results regarding white matter (WM) integrity in HEU children, often varying by geography and methodology. Furthermore, no research had previously investigated the relationship between auditory WM tract development and language outcomes specifically in iHEU infants. The study aims to determine if in utero HIV and ART exposure alters the microstructural integrity of auditory WM tracts in newborns and whether these structural changes predict language development at 9–14 months.

2. Methodology

• Study Cohort:

  • Source: The Healthy Baby Study (HBS) in South Africa.
  • Participants: 85 infants scanned at 0–5 weeks of age.
    • iHEU-pre (n=31): Mothers initiated ART pre-conception.
    • iHEU-post (n=29): Mothers initiated ART post-conception (second trimester).
    • iHU (n=25): HIV-unexposed controls.
  • Follow-up: 55 infants underwent language assessments at 9–14 months.

• Imaging Protocol:

  • Scanner: 3T Siemens Skyra.
  • Sequence: Diffusion Tensor Imaging (DTI) using a multi-band spin-echo EPI sequence (30 directions, b=1000 s/mm²).
  • Acquisition: Natural sleep (no sedation); motion correction via volumetric navigators.

• Image Processing & Tractography:

  • Preprocessing: Infant FreeSurfer for T1 segmentation; TORTOISE (DIFF_PREP, DR-BUDDI) for motion/eddy current correction and distortion correction.
  • Regions of Interest (ROIs):
    • 28 automated structures (e.g., thalamus, cerebellum, ventricles).
    • 8 manually traced auditory structures: Cochlear Nuclei (CN), Inferior Colliculus (IC), Medial Geniculate Nucleus (MGN), and Heschl's Gyrus (HG).
  • Tractography: Full-probabilistic tractography (3dTrackID) to identify connections between all seed pairs.
    • Whole-brain: Angle threshold 60°, min length 15mm, FA threshold 0.1.
    • Intra-auditory: Optimized thresholds (max angle 55°, min length 10mm).
  • Analysis: 106 auditory connections (involving at least one auditory seed) were analyzed.

• Statistical Analysis:

  • Tools: R software.
  • Models: Linear regression to compare DTI metrics (Fractional Anisotropy [FA], Mean Diffusivity [MD], Radial Diffusivity [RD], Axial Diffusivity [AD]) between groups, controlling for confounders (maternal education, weight gain).
  • Corrections: False Discovery Rate (FDR) for multiple comparisons (q ≤ 0.05); unadjusted p-values (p ≤ 0.05) reported for exploratory patterns.
  • Effect Size: Cohen's d and $r^2$ calculated to compare findings with existing literature (Magondo et al., 2024; Liu et al., 2012).
  • Language Assessment: Griffiths Mental Development Scales (GMDS) and Griffiths III (Language Quotient).

3. Key Contributions

1. First Investigation of Central Auditory Tracts in iHEU: This is the first study to use probabilistic tractography to map specific auditory WM pathways (including brainstem and thalamic connections) in iHEU infants.
2. Differentiation of ART Timing: The study explicitly separates the effects of early (pre-conception) vs. late (post-conception) ART exposure, addressing whether longer exposure duration offers neuroprotection or increased risk.
3. Structure-Function Correlation: It investigates the link between neonatal auditory tract integrity and later language outcomes, revealing group-specific correlations (present in controls, absent in iHEU).
4. Effect Size Contextualization: By calculating Cohen's d and comparing it to other infant DTI studies, the authors provide a benchmark for the clinical meaningfulness of subtle DTI changes in this population, even when statistical significance is lost after multiple comparison correction.

4. Key Results

• Statistical Significance: No results remained statistically significant after FDR correction for multiple comparisons.
• Unadjusted Findings (FA):
  • iHEU infants showed reduced Fractional Anisotropy (FA) in 6 auditory connections compared to iHU.
  • This reduction was primarily driven by the iHEU-post group.
  • Affected tracts connected auditory structures (e.g., MGN) to non-auditory structures (e.g., cerebral cortex, thalamus).
• Unadjusted Findings (MD):
  • iHEU infants showed elevated Mean Diffusivity (MD) in 22 auditory connections.
  • Both iHEU-pre and iHEU-post contributed to these elevations, suggesting the effect is not solely dependent on ART duration.
  • Left Hemisphere Bias: The Left Medial Geniculate Nucleus (L MGN) was the most frequently affected structure (41% of affected tracts).
  • Elevated MD was often accompanied by elevated RD and AD, suggesting delayed myelination and fiber organization.
• Language Outcomes:
  • No Group Differences: There were no significant differences in GMDS or Griffiths III language scores between iHEU and iHU groups at 9–14 months.
• DTI-Language Correlations:
  • iHU Group: Showed moderately strong correlations between DTI metrics and language scores.
    • Positive correlation: Higher FA in tracts connected to the Cochlear Nucleus (CN) correlated with better language.
    • Negative correlation: Lower MD in tracts connected to the L MGN and CN correlated with better language.
  • iHEU Group: No meaningful correlations were found between DTI metrics and language scores in the affected tracts.
• Effect Sizes: The observed effect sizes (Cohen's d ~0.6–0.8 for specific tracts) were comparable to or larger than those reported in other infant DTI studies (e.g., Magondo et al., 2024), despite the lack of statistical significance after correction.

5. Significance and Conclusion

• Subtle Neurodevelopmental Impact: The study suggests that HIV/ART exposure may cause subtle delays in auditory WM maturation, particularly in tracts connecting to the left auditory thalamus (MGN). These delays appear to be driven by HIV exposure itself rather than the duration of ART, as both pre- and post-exposure groups showed MD elevations.
• Decoupling of Structure and Function: While iHEU infants showed structural alterations (reduced FA, elevated MD), these did not translate to detectable language deficits at 9–14 months. However, the absence of structure-function correlations in iHEU (unlike in iHU) suggests that the mechanism of maturation in auditory tracts may be fundamentally altered in iHEU infants.
• Clinical Implications: The findings imply that while language performance appears normal in early infancy, the underlying neural architecture supporting auditory processing may be developing differently. This warrants long-term monitoring to see if these subtle structural differences manifest as language or auditory processing deficits later in childhood.
• Limitations: The study was limited by a small sample size (reducing power after multiple comparison correction) and the inherent difficulty of imaging small brainstem structures in neonates. Future work requires larger cohorts and along-tract analysis to pinpoint specific developmental disruptions.

In summary, the paper provides evidence that HIV exposure, independent of ART duration, may subtly disrupt the microstructural development of central auditory pathways in newborns, potentially altering the trajectory of how these tracts support language processing, even if overt language delays are not yet apparent.