So, you’ve taken your neurological history and completed the impossibly complex neurological examination and you finally reached some sort of conclusion as to what you think is wrong with the patient.
Your next challenge, or nightmare, depending on your point of view is to try and come up with a sensible plan to investigate the person in front of you without resorting to the most bizarre and niche test available on the National Health Service.
Obviously, you want to start with the most basic things available and then work your way up. Not only is this sensible for your exams, but also stops you missing things when you qualify as a doctor.
I constantly have to remind myself to do this even has a consultant.
Without wishing to be crude, I always start with the question: “What is the easiest thing to take out of this person?” Usually, the easiest thing to take out of anybody is to take the “piss” out of them. So your starting point for any investigation plan should be urinalysis
This is how I usually structure my thoughts.
Really easy things to take out
- urine
- sputum
- stool
- puff (see below)
Pretty easy things to take out
- venous blood
- arterial blood
Harder things to take out
- CSF
- biopsies
The next step is to think about imaging. Again, try and start with the most basic form of medical imaging and work your way up.
Basic imaging
- ECG
- CXR
Fancier imaging
- CT
- MRI
Really fancy stuff
- EEG
- Nerve conduction studies
- Radioisotope scans
Let’s look at the these in a little more detail.
Stuff you take out
Urine
There probably aren’t that many situations where the urinalysis is a game changer in a neurological presentation. It is always helpful to know that your patient doesn’t have incipient urinary sepsis, particularly given that we use a lot of immunosuppression. Similarly, I have lost count of the number of times I have had to go back and request a pregnancy test before starting a new drug, wishing I had thought to do it earlier.
Confused or drowsy patients may have taken drugs so a urine toxicology screen is always a good option (and you will probably be the only one to think to do it on admission).
Sputum
See above about sepsis. I hope there are better ways to pick up TB now, but I might be wrong. It is a great neurological mimic so if there is a cough with sputum, always worth a bit of a look.
Stool
I became a neurologist to avoid this kind of thing so I’m definitely not dwelling on it. You can do your own work on this topic!
Puff
Now this is a good one to remember not to forget. Loads of Neuro patients have respiratory problems – Guillain-Barré syndrome, myasthenia, MND to name but a few.
Having a good idea of their lung function is really important.
Whatever you do, don’t do peak flow. This is a terrible test for neuromuscular weakness. Respiratory rate and saturations are fine, up to a point, but sats drop really late, like nearly dead late, and respiratory rate is unreliable when patients are fatigued, weak and tired.
You really need to know how the diaphragm is working and, as such, the Forced Vital Capacity (FVC) is much more useful.

It can be a nightmare trying to do this in a neuro patient, for two main reasons – firstly, you can never get a machine to measure it and, secondly, if they are weak in the face they might not be able to make a seal and perform the test.
You can do something called a breath count. Just ask the patient to take a deep breath in and then start counting out loud in a normal volume voice, so you can hear them. Tell them to count as far as they can in one breath. Try it yourself and see how far you get – 30? 40? More?
In it’s own right this information is not that helpful, as it will vary in each person depending on how loud they speak, how fast they count, how big they are etc. etc. But, if you get a baseline for that person, say “28”, and then 4 hours later it is “21” and four hours after that it is “13”, then you know their vital capacity is dropping and you need help from HDU/ICU.
Venous blood
This is a bit of a cheat. I only stress the “venous” bit to remind myself that “arterial” blood exists. You would be amazed how many times folk forget to do a blood gas when an ill patient. I reckon if you make yourself think of venous and arterial at the start of a plan of investigation it makes this omission less likely.
You can send off any number of blood tests of course. It would bore me, and you, to go over them in detail but, basically, think of it like this:
- Haematology – including odd little things like ESR, blood films
- Biochemistry – routine, CRP and fancier bits and bobs
- Endocrine – glucose, cortisol, other hormones
- Microbiology – cultures, PCRs
- Immunology – autoimmune screens
- Others
Arterial blood gases
These are really great when patients are critically ill, and mostly they get done really quickly in acute care settings. You get all the key parameters and a quick idea of electrolytes too.
I’m a big fan in patients with potential neuromuscular weakness, where the other respiratory tests ring all the right alarms but far too late in the day.
Look out for a rising pCO2 first, long before the O2 levels drop. Try to avoid too much oxygen in this type of patient as, like someone with COPD, you can kill them with kindness by removing the hypoxic drive.
CSF
Now this is one of my favourite tests, and not just because I am a neurologist. For starters, it is usually a lovely clear colour – not like that nasty mucus stuff and it is pleasantly odourless. When it isn’t clear there is serious mischief afoot and so your intervention is well worth it in any case.
It’s also lovely to be absorbed in the procedure for 20-30 minutes – no phone, no bleep, no interruptions. Relative Zen calm descends on the LPer.
There are a number of presentations amenable to CSF analysis. For our purposes, we are most interested in the following:
- acute headache – particularly if you are looking for blood or blood products
- CNS infection – viral, bacterial, fungal, parasites
- CNS immune disease (like MS) – usually we are looking for immune products (called oligoclonal bands)
- potential raised pressure – what I mean is a condition called Idiopathic Intracranial Hypertension, where CSF pressure is raised in otherwise healthy, if overweight, people
There are loads of other instances but this isn’t meant to be a definitive guide.
CSF analysis
Opening pressure – this should always be measured although often isn’t as people can’t find a manometer to attach. It should be < 26 mm CSF (and definitely < 30). It is raised in venous sinus thrombosis (DVT in the head) and can be the only clue sometimes. It is also raised in IIH, infections are other conditions too.
RBC – raised if you prang a vessel on the way in (common) or in subarachnoid haemorrhage. Counting blood cells is a mugs game – don’t do it. Blood is broken down after 12 hours into xanthochromia. These hang about for about 14 days so, if someone presents with a possible SAH, and their scan is normal, xanthochromia analysis is usually the next step. Don’t do the LP too soon.
WBC – loads of things put up the white cells. Too many red cells = too many white cells so it is hard to interpret in a traumatic tap. In general, immune conditions raise the WCC a little bit (10-50) and infection raises it a lot (100-1000). Lymphocytes suggest viral infection; Neutrophils suggest bacterial infection
Protein – not very specific and is raised in all sorts of conditions. Normally, it should be less than 0.6 g/L. If it is over 1.0 g/L something is going on!
Glucose – should be about 50% of serum levels (so you need to take a SIMLUTANEOUS blood sample). Cancer cells and bacteria (particularly TB, meningococcus) eat glucose for breakfast so a low CSF glucose (< 30% serum) usually suggests serious mischief. To be fair, the rest of the CSF parameters are wonky as well in this situation.
Microscopy & Culture – always send a nice sample to microbiology for microscopy, culture and PCR
Cytology – lymphoma and cancer can secrete into the CSF but it can take several CSF studies to pick them up.
Immune state – MS is a CSF-specific immune disorder and so you see antibodies (oligoclonal bands) in the CSF of most patients (>90%). Conversely, you do not see these in the serum in a condition like MS. If there are bands in the CSF, but not in the serum, we call these “unmatched” and, in the right clinical circumstances, can be helpful in diagnosis of MS. OCBs in both serum and CSF don’t really tell you much.
Basic imaging
ECG
This might be my favourite neurological test. It is certainly the cheapest. No patient with a blackout should escape your clutches without an ECG. We refuse referrals if the ECG hasn’t been done.
Cardiac problems can mimic seizures and even the most experienced neurologists will have missed structural heart disease or cardiac arrhythmias.
Of particular note is the long QT syndrome.
You might need to build on this, with prolonged monitoring, implantable loop recorders or ECHO tests but and ECG is the starting point.
CXR
This is the second cheapest and you would be amazed how many times a patient has had loads of fancy tests but no CXR. All sorts of things can be picked up – heart disease, lung cancer, sarcoid. Well worth a shout in your investigation scheme.

Fancier imaging
CT
Much overlooked now, but I still have a big soft spot for CT scans.
For starts, they are quick and easy to tolerate for patients. They are also pretty straightforward to arrange.
In the first 24 hours for an acute headache presentation, they show up > 90% of subarachnoid haemorrhage. A normal CT doesn’t rule out SAH so one often needs the LP as well. A positive CT for SAH necessitates much more careful neurosurgical assessment.
You can also see the arterial and venous system really well – if you ask your radiologists and radiographers in advance.
Contrast is needed so aneurysms and venous sinus thrombosis both show up well on CT angio and CT venogram.

Contrast can also light up tumours and other nasties that you might otherwise miss.
Before you request a CT, take a moment to think if contrast might help and if so, add it too the request.
We often do a lot of CT thorax, abdomen and pelvis scans as well, looking for hidden tumours in patients with paraneoplastic disorders. Not your first test, of course, as it is a fair old dose of radiation, but can be vital in the right setting.
MRI
Now, I’m no expert, but there is a reason why this got the Nobel prize in Physiology and Medicine in 2003. It is quite frankly as life saver – not least for the stranded neurologist!
I’m definitely not going to tell you how it works. It doesn’t really matter for our purposes.
What is useful to know is that there are a number of scans within the scan, if that makes sense. Some are good for anatomy, some are good for pathology, some are good for conditions (stroke, haemorrhage, MS, tumours etc) so what you request, and what you get, depends on the information you give the radiologists and radiographers – so DO NOT SKIMP ON THE REQUEST DETAILS PLEASE.
You don’t need to memorise this but, just for completeness:
T1 – good for structure and degrees of atrophy
T2 – good for “pathology” – like MS lesions, white matter disease
Flair – really good for inflammatory lesions, MS, tumours etc.

DWI/ADC – fantastic for ischaemic stroke

Contrast – lights up tumours, abscesses and recent MS plaques
Angiography/Venography – shows up aneurysms and the venous drainage system.

Radioisotope scans
Really not core knowledge but I sometimes request an isotope scan of the dopamine system (dopamine transported scan) when investigating some patients with tremor and subtle Parkinsonism where I’m not sure what the diagnosis is.
Really fancy stuff
EEG
I’m not an expert in this at all and my colleagues in the neurophysiology department will despair if I try and cover it in any detail.
Basically, EEG can be helpful in the following situations:
- young patients with seizures (children and young adults) – who may have a genetic epilepsy syndrome
- patients with dissociative/non-epileptic attacks – who may have a typical attack during an EEG and allow you to diagnose their episodes with more confidence
- high-frequency seizures where careful localisation of the seizure onset zone may allow for surgical resection
- questions of ongoing seizure activity, despite sedation, in a patient on ICU

There is no role for EEG in the following situations:
- unexplained blackout ?cause
- episodes of rage in psychiatric units
- patients in whom your diagnosis of seizures is clinically secure
Far too many EEGs are requested and, usually, they are as much use as a chocolate teapot. They do not help if you are not sure what the blackout/attack is caused by and should not be used as a fishing expedition when in search of certainty. Instead, embrace uncertainty and live with it!
Nerve conduction studies
These are wonderful things, full of diagnostic insight and performed by people with an intimate knowledge of peripheral neuroanatomy and neurophysiology. This is why I have no clue how to do or interpret them!
Basically, we can use peripheral neurophysiology to look at the sensory and motor nerves for conditions like neuropathy and motor neurone disease. Myasthenia gravis, a disorder of the neuromuscular junction has its own signature changes as well.
